docs(brand): add hero banner

Co-authored-by: zeekay <z@hanzo.ai>

.env.example

@@ -1,11 +1,11 @@
-# PostHog API Configuration
+# Hanzo Insights API Configuration
 # Copy this file to .env and update with your actual values
 
-# Your project API key (found on the /setup page in PostHog)
-POSTHOG_PROJECT_API_KEY=phc_your_project_api_key_here
+# Your project API key (found on the setup page in Insights)
+INSIGHTS_PROJECT_API_KEY=hi_your_project_api_key_here
 
 # Your personal API key (for local evaluation and other advanced features)
-POSTHOG_PERSONAL_API_KEY=phx_your_personal_api_key_here
+INSIGHTS_PERSONAL_API_KEY=phx_your_personal_api_key_here
 
-# PostHog host URL (remove this line if using posthog.com)
-POSTHOG_HOST=http://localhost:8000
+# Insights host URL (remove this line if using insights.hanzo.ai)
+INSIGHTS_HOST=http://localhost:8000

.github/hero.svg

@@ -0,0 +1,9 @@
+<svg xmlns="http://www.w3.org/2000/svg" width="1280" height="640" viewBox="0 0 1280 640" role="img" aria-label="insights-python">
+<rect width="1280" height="640" fill="#0A0A0A"/>
+<svg x="96" y="215" width="210" height="210" viewBox="0 0 67 67"><path d="M22.21 67V44.6369H0V67H22.21Z" fill="#fff"/><path d="M66.7038 22.3184H22.2534L0.0878906 44.6367H44.4634L66.7038 22.3184Z" fill="#fff"/><path d="M22.21 0H0V22.3184H22.21V0Z" fill="#fff"/><path d="M66.7198 0H44.5098V22.3184H66.7198V0Z" fill="#fff"/><path d="M66.7198 67V44.6369H44.5098V67H66.7198Z" fill="#fff"/></svg>
+<text x="378" y="276" font-family="Inter,system-ui,-apple-system,sans-serif" font-size="78" font-weight="800" letter-spacing="-2" fill="#ffffff">insights-python</text>
+<text x="378" y="322" font-family="Inter,system-ui,sans-serif" font-size="30" fill="#ffffff" opacity=".66">Send usage data from your Python code to PostHog.</text>
+<rect x="378" y="338" width="806" height="3" rx="1.5" fill="#ffffff" opacity=".9"/>
+<text x="378" y="390" font-family="Inter,system-ui,sans-serif" font-size="24" font-weight="600" fill="#ffffff" opacity=".5">github.com/hanzoai</text>
+<text x="1184" y="390" text-anchor="end" font-family="Inter,system-ui,sans-serif" font-size="24" font-weight="600" fill="#ffffff" opacity=".5">hanzo.ai</text>
+</svg>

.github/workflows/ci.yml

@@ -9,7 +9,7 @@ permissions:
 jobs:
     code-quality:
         name: Code quality checks
-        runs-on: ubuntu-latest
+        runs-on: hanzo-build-linux-amd64
 
         steps:
             - uses: actions/checkout@85e6279cec87321a52edac9c87bce653a07cf6c2
@@ -46,7 +46,7 @@ jobs:
 
     tests:
         name: Python ${{ matrix.python-version }} tests
-        runs-on: ubuntu-latest
+        runs-on: hanzo-build-linux-amd64
         strategy:
             matrix:
                 python-version: ['3.10', '3.11', '3.12', '3.13', '3.14']
@@ -76,9 +76,32 @@ jobs:
               run: |
                   pytest --verbose --timeout=30
 
+    import-check:
+        name: Python ${{ matrix.python-version }} import check
+        runs-on: hanzo-build-linux-amd64
+        strategy:
+            matrix:
+                python-version: ['3.10', '3.11', '3.12', '3.13', '3.14']
+
+        steps:
+            - uses: actions/checkout@85e6279cec87321a52edac9c87bce653a07cf6c2
+              with:
+                  fetch-depth: 1
+
+            - name: Set up Python ${{ matrix.python-version }}
+              uses: actions/setup-python@8d9ed9ac5c53483de85588cdf95a591a75ab9f55
+              with:
+                  python-version: ${{ matrix.python-version }}
+
+            - name: Install posthog
+              run: pip install .
+
+            - name: Check import produces no warnings
+              run: python -W error -c "import posthog"
+
     django5-integration:
         name: Django 5 integration tests
-        runs-on: ubuntu-latest
+        runs-on: hanzo-build-linux-amd64
 
         steps:
             - uses: actions/checkout@85e6279cec87321a52edac9c87bce653a07cf6c2

.github/workflows/codeql.yml

@@ -0,0 +1,45 @@
+name: 'CodeQL Advanced'
+
+on:
+  push:
+    branches: ['master']
+  pull_request:
+    branches: ['master']
+  schedule:
+    - cron: '32 13 * * 1'
+
+jobs:
+  analyze:
+    name: Analyze (${{ matrix.language }})
+    runs-on: hanzo-build-linux-amd64
+    permissions:
+      security-events: write
+      # required to fetch internal or private CodeQL packs
+      packages: read
+
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - language: actions
+            build-mode: none
+          - language: python
+            build-mode: none
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v6
+
+      - name: Initialize CodeQL
+        uses: github/codeql-action/init@5d4e8d1aca955e8d8589aabd499c5cae939e33c7 # v4.31.9
+        with:
+          languages: ${{ matrix.language }}
+          build-mode: ${{ matrix.build-mode }}
+          # Disable TRAP caching - it creates a new cache per commit SHA which
+          # is never reused, causing wasted cache space.
+          # See: https://github.com/github/codeql-action/issues/2030
+          trap-caching: false
+
+      - name: Perform CodeQL Analysis
+        uses: github/codeql-action/analyze@5d4e8d1aca955e8d8589aabd499c5cae939e33c7 # v4.31.9
+        with:
+          category: '/language:${{matrix.language}}'

.github/workflows/generate-references.yml

@@ -6,7 +6,7 @@ on:
 jobs:
   docs-generation:
     name: Generate references
-    runs-on: ubuntu-latest
+    runs-on: hanzo-build-linux-amd64
     permissions:
       contents: write
     steps:

.github/workflows/release.yml

@@ -1,56 +1,249 @@
 name: "Release"
 
 on:
-  push:
-    branches:
-      - master
-    paths:
-      - "posthog/version.py"
+  pull_request:
+    types: [closed]
+    branches: [master]
   workflow_dispatch:
 
+permissions:
+  contents: read
+
+# Concurrency control: only one release process can run at a time
+# This prevents race conditions if multiple PRs with 'release' label merge simultaneously
+concurrency:
+  group: release
+  cancel-in-progress: false
+
 jobs:
-  release:
-    name: Publish release
-    runs-on: ubuntu-latest
-    permissions:
-      contents: write
-      id-token: write
+  check-release-label:
+    name: Check for release label
+    runs-on: hanzo-build-linux-amd64
+    # Run when PR with 'release' label is merged to master
+    if: |
+      github.event_name == 'workflow_dispatch' ||
+       (github.event_name == 'pull_request' &&
+        github.event.pull_request.merged == true &&
+        contains(github.event.pull_request.labels.*.name, 'release'))
+    outputs:
+      should-release: ${{ steps.check.outputs.should-release }}
     steps:
-      - name: Checkout the repository
-        uses: actions/checkout@85e6279cec87321a52edac9c87bce653a07cf6c2
+      - name: Checkout repository
+        uses: actions/checkout@v4
         with:
+          ref: master
           fetch-depth: 0
 
+      - name: Check release conditions
+        id: check
+        run: |
+          changeset_count=$(find .sampo/changesets -name '*.md' 2>/dev/null | wc -l)
+          if [ "$changeset_count" -gt 0 ]; then
+            echo "should-release=true" >> "$GITHUB_OUTPUT"
+            echo "Found $changeset_count changeset(s), ready to release"
+          else
+            echo "should-release=false" >> "$GITHUB_OUTPUT"
+            echo "No changesets to release"
+          fi
+
+  notify-approval-needed:
+    name: Notify Slack - Approval Needed
+    needs: check-release-label
+    if: needs.check-release-label.outputs.should-release == 'true'
+    uses: posthog/.github/.github/workflows/notify-approval-needed.yml@main
+    with:
+      slack_channel_id: ${{ vars.SLACK_APPROVALS_CLIENT_LIBRARIES_CHANNEL_ID }}
+      slack_user_group_id: ${{ vars.GROUP_CLIENT_LIBRARIES_SLACK_GROUP_ID }}
+    secrets:
+      slack_bot_token: ${{ secrets.SLACK_CLIENT_LIBRARIES_BOT_TOKEN }}
+      posthog_project_api_key: ${{ secrets.POSTHOG_PROJECT_API_KEY }}
+
+  release:
+    name: Release and publish
+    needs: [check-release-label, notify-approval-needed]
+    runs-on: hanzo-build-linux-amd64
+    # Use `always()` to ensure the job runs even if notify-approval-needed is skipped,
+    # but still depend on it to access `needs.notify-approval-needed.outputs.slack_ts`
+    if: always() && needs.check-release-label.outputs.should-release == 'true'
+    environment: "Release" # This will require an approval from a maintainer, they are notified in Slack above
+    permissions:
+      contents: write
+      actions: write
+      id-token: write
+    steps:
+      - name: Notify Slack - Approved
+        if: needs.notify-approval-needed.outputs.slack_ts != ''
+        uses: posthog/.github/.github/actions/slack-thread-reply@main
+        with:
+          slack_bot_token: ${{ secrets.SLACK_CLIENT_LIBRARIES_BOT_TOKEN }}
+          slack_channel_id: ${{ vars.SLACK_APPROVALS_CLIENT_LIBRARIES_CHANNEL_ID }}
+          thread_ts: ${{ needs.notify-approval-needed.outputs.slack_ts }}
+          message: "✅ Release approved! Version bump in progress..."
+          emoji_reaction: "white_check_mark"
+
+      - name: Get GitHub App token
+        id: releaser
+        uses: actions/create-github-app-token@v2
+        with:
+          app-id: ${{ secrets.GH_APP_POSTHOG_PYTHON_RELEASER_APP_ID }}
+          private-key: ${{ secrets.GH_APP_POSTHOG_PYTHON_RELEASER_PRIVATE_KEY }}
+
+      - name: Checkout repository
+        uses: actions/checkout@v4
+        with:
+          ref: master
+          fetch-depth: 0
+          token: ${{ steps.releaser.outputs.token }}
+
       - name: Set up Python
-        uses: actions/setup-python@8d9ed9ac5c53483de85588cdf95a591a75ab9f55
+        uses: actions/setup-python@v5
         with:
           python-version: 3.11.11
 
       - name: Install uv
-        uses: astral-sh/setup-uv@0c5e2b8115b80b4c7c5ddf6ffdd634974642d182 # v5.4.1
+        uses: astral-sh/setup-uv@v5
         with:
-            enable-cache: true
-            pyproject-file: 'pyproject.toml'
-            
-      - name: Detect version
-        run: echo "REPO_VERSION=$(python3 posthog/version.py)" >> $GITHUB_ENV
+          enable-cache: true
+          pyproject-file: "pyproject.toml"
 
-      - name: Prepare for building release
+      - name: Install Rust
+        uses: dtolnay/rust-toolchain@0b1efabc08b657293548b77fb76cc02d26091c7e
+        with:
+          toolchain: 1.91.1
+          components: cargo
+
+      - name: Cache Sampo CLI
+        id: cache-sampo
+        uses: actions/cache@v3
+        with:
+          path: ~/.cargo/bin/sampo
+          key: sampo-${{ runner.os }}-${{ runner.arch }}
+
+      - name: Install Sampo CLI
+        if: steps.cache-sampo.outputs.cache-hit != 'true'
+        run: cargo install sampo
+
+      - name: Install dependencies
         run: uv sync --extra dev
 
-      - name: Push releases to PyPI
-        env:
-          TWINE_USERNAME: __token__
-        run: uv run make release && uv run make release_analytics
+      - name: Configure Git
+        run: |
+          git config user.name "github-actions[bot]"
+          git config user.email "github-actions[bot]@users.noreply.github.com"
 
-      - name: Create GitHub release
-        uses: actions/create-release@0cb9c9b65d5d1901c1f53e5e66eaf4afd303e70e # v1
-        with:
-          tag_name: v${{ env.REPO_VERSION }}
-          release_name: ${{ env.REPO_VERSION }}
-      
-      - name: Dispatch generate-references for posthog-python
+      - name: Prepare release with Sampo
+        id: sampo-release
+        env:
+          GITHUB_TOKEN: ${{ steps.releaser.outputs.token }}
+        run: |
+          sampo release
+          new_version=$(python3 -c "import tomllib; print(tomllib.load(open('pyproject.toml', 'rb'))['project']['version'])")
+          echo "new_version=$new_version" >> "$GITHUB_OUTPUT"
+
+      - name: Sync version to posthog/version.py
+        run: |
+          echo 'VERSION = "${{ steps.sampo-release.outputs.new_version }}"' > posthog/version.py
+
+      - name: Commit release changes
+        id: commit-release
+        env:
+          GITHUB_TOKEN: ${{ steps.releaser.outputs.token }}
+        run: |
+          git add -A
+          if git diff --staged --quiet; then
+            echo "No changes to commit"
+            echo "committed=false" >> "$GITHUB_OUTPUT"
+          else
+            git commit -m "chore: Release v${{ steps.sampo-release.outputs.new_version }}"
+            git push origin master
+            echo "committed=true" >> "$GITHUB_OUTPUT"
+          fi
+
+      # Publishing is done manually (not via `sampo publish`) because we need to
+      # publish both `posthog` and `posthoganalytics` packages to PyPI.
+      # Sampo only knows about the `posthog` package, so we handle both here.
+      # Both packages use PyPI OIDC trusted publishing (no API tokens needed).
+      - name: Build posthog
+        if: steps.commit-release.outputs.committed == 'true'
+        run: uv run make build_release
+
+      - name: Publish posthog to PyPI
+        if: steps.commit-release.outputs.committed == 'true'
+        uses: pypa/gh-action-pypi-publish@release/v1
+
+      # The `posthoganalytics` package is a mirror of `posthog` published under
+      # a different name for backwards compatibility. The make target handles
+      # copying, renaming imports, and building the dist automatically.
+      - name: Build posthoganalytics
+        if: steps.commit-release.outputs.committed == 'true'
+        run: uv run make build_release_analytics
+
+      - name: Publish posthoganalytics to PyPI
+        if: steps.commit-release.outputs.committed == 'true'
+        uses: pypa/gh-action-pypi-publish@release/v1
+
+      # We skip `sampo publish` (which normally creates the tag) because we
+      # need to publish both posthog and posthoganalytics manually, so we
+      # create the tag ourselves.
+      - name: Tag release
+        if: steps.commit-release.outputs.committed == 'true'
+        run: git tag "v${{ steps.sampo-release.outputs.new_version }}"
+
+      - name: Push tags
+        if: steps.commit-release.outputs.committed == 'true'
+        run: git push origin --tags
+
+      - name: Create GitHub Release
+        if: steps.commit-release.outputs.committed == 'true'
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        run: gh release create "v${{ steps.sampo-release.outputs.new_version }}" --generate-notes
+
+      - name: Dispatch generate-references
+        if: steps.commit-release.outputs.committed == 'true'
         env:
           GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        run: |
-          gh workflow run generate-references.yml --ref master
+        run: gh workflow run generate-references.yml --ref master
+
+      # Notify in case of a failure
+      - name: Send failure event to PostHog
+        if: ${{ failure() }}
+        uses: PostHog/posthog-github-action@v0.1
+        with:
+          posthog-token: "${{ secrets.POSTHOG_PROJECT_API_KEY }}"
+          event: "posthog-python-github-release-workflow-failure"
+          properties: >-
+            {
+              "commitSha": "${{ github.sha }}",
+              "jobStatus": "${{ job.status }}",
+              "ref": "${{ github.ref }}",
+              "version": "v${{ steps.sampo-release.outputs.new_version }}"
+            }
+
+      - name: Notify Slack - Failed
+        if: ${{ failure() && needs.notify-approval-needed.outputs.slack_ts != '' }}
+        uses: posthog/.github/.github/actions/slack-thread-reply@main
+        with:
+          slack_bot_token: ${{ secrets.SLACK_CLIENT_LIBRARIES_BOT_TOKEN }}
+          slack_channel_id: ${{ vars.SLACK_APPROVALS_CLIENT_LIBRARIES_CHANNEL_ID }}
+          thread_ts: ${{ needs.notify-approval-needed.outputs.slack_ts }}
+          message: "❌ Failed to release `posthog-python@v${{ steps.sampo-release.outputs.new_version }}`! <https://github.com/${{ github.repository }}/actions/runs/${{ github.run_id }}|View logs>"
+          emoji_reaction: "x"
+
+  notify-released:
+    name: Notify Slack - Released
+    needs: [check-release-label, notify-approval-needed, release]
+    runs-on: hanzo-build-linux-amd64
+    if: always() && needs.release.result == 'success' && needs.notify-approval-needed.outputs.slack_ts != ''
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+
+      - name: Notify Slack - Released
+        uses: posthog/.github/.github/actions/slack-thread-reply@main
+        with:
+          slack_bot_token: ${{ secrets.SLACK_CLIENT_LIBRARIES_BOT_TOKEN }}
+          slack_channel_id: ${{ vars.SLACK_APPROVALS_CLIENT_LIBRARIES_CHANNEL_ID }}
+          thread_ts: ${{ needs.notify-approval-needed.outputs.slack_ts }}
+          message: "🚀 posthog-python released successfully!"
+          emoji_reaction: "rocket"

.github/workflows/sdk-compliance.yml

@@ -0,0 +1,21 @@
+name: SDK Compliance Tests
+
+permissions:
+  contents: read
+  packages: read
+  pull-requests: write
+
+on:
+  pull_request:
+  push:
+    branches:
+      - master
+
+jobs:
+  compliance:
+    name: PostHog SDK compliance tests
+    uses: PostHog/posthog-sdk-test-harness/.github/workflows/test-sdk-action.yml@main
+    with:
+      adapter-dockerfile: "sdk_compliance_adapter/Dockerfile"
+      adapter-context: "."
+      test-harness-version: "latest"

.sampo/changesets/prompt-sdk-versions.md

@@ -0,0 +1,5 @@
+---
+pypi/posthog: patch
+---
+
+feat(llma): support fetching versioned prompts from the prompts sdk

.sampo/config.toml

@@ -0,0 +1,19 @@
+# Sampo configuration
+version = 1
+
+[git]
+default_branch = "master"
+short_tags = "posthog"    # Tag with v1.2.3 rather than posthog-v1.2.3
+
+[github]
+repository = "posthog/posthog-python"
+
+[changelog]
+# Options for release notes generation.
+# show_commit_hash = true (default)
+# show_acknowledgments = true (default)
+
+[packages]
+# Options for package discovery and filtering.
+# ignore_unpublished = false (default)
+# ignore = ["internal-*", "examples/*"]

BEFORE_SEND.md

@@ -1,6 +1,6 @@
 # Before Send Hook
 
-The `before_send` parameter allows you to modify or filter events before they are sent to PostHog. This is useful for:
+The `before_send` parameter allows you to modify or filter events before they are sent to Insights. This is useful for:
 
 - **Privacy**: Removing or masking sensitive data (PII)
 - **Filtering**: Dropping unwanted events (test events, internal users, etc.)
@@ -10,12 +10,12 @@ The `before_send` parameter allows you to modify or filter events before they ar
 ## Basic Usage
 
 ```python
-import posthog
+import hanzo_insights
 from typing import Optional, Dict, Any
 
 def my_before_send(event: Dict[str, Any]) -> Optional[Dict[str, Any]]:
     """
-    Process event before sending to PostHog.
+    Process event before sending to Insights.
     
     Args:
         event: The event dictionary containing 'event', 'distinct_id', 'properties', etc.
@@ -27,7 +27,7 @@ def my_before_send(event: Dict[str, Any]) -> Optional[Dict[str, Any]]:
     return event
 
 # Initialize client with before_send hook
-client = posthog.Client(
+client = hanzo_insights.Client(
     api_key="your-project-api-key",
     before_send=my_before_send
 )
@@ -166,7 +166,7 @@ def should_drop_event(event: dict[str, Any]) -> bool:
 
 ## Error Handling
 
-If your `before_send` function raises an exception, PostHog will:
+If your `before_send` function raises an exception, Insights will:
 
 1. Log the error 
 2. Continue with the original, unmodified event
@@ -184,7 +184,7 @@ def risky_before_send(event: dict[str, Any]) -> Optional[dict[str, Any]]:
 ## Complete Example
 
 ```python
-import posthog
+import hanzo_insights
 from typing import Optional, Any
 import re
 
@@ -227,7 +227,7 @@ def production_before_send(event: dict[str, Any]) -> Optional[dict[str, Any]]:
         return event  # Return original event on error
 
 # Usage
-client = posthog.Client(
+client = hanzo_insights.Client(
     api_key="your-api-key",
     before_send=production_before_send
 )

CHANGELOG.md

@@ -1,18 +1,116 @@
-# 7.4.3 - 2026-01-02
+# posthog
+
+## 7.9.7 — 2026-03-05
+
+### Patch changes
+
+- [b206669](https://github.com/posthog/posthog-python/commit/b206669bf62c923346ad28881dc4694d933ca424) fix(llma): use distinct_id from outer context if not provided, fix $process_person_profile for context-based identity — Thanks @ethanporcaro for your first contribution 🎉!
+- [a99c7d7](https://github.com/posthog/posthog-python/commit/a99c7d73b1e0ef1f35d856c82ace21237ee253a3) Add warning log for local flag evaluation cold start — Thanks @dmarticus!
+
+## 7.9.6 — 2026-03-02
+
+### Patch changes
+
+- [8d83315](https://github.com/posthog/posthog-python/commit/8d83315b67c21eb9e7d6c17bae27ada98ca2643d) add PROPERTY_OPERATORS constant for match_property — Thanks @dmarticus!
+
+## 7.9.5 — 2026-03-02
+
+### Patch changes
+
+- [830244b](https://github.com/posthog/posthog-python/commit/830244bd409b1992ae2e49610f8f87d2cdfc8096) add semver targeting support to local evaluation — Thanks @dmarticus!
+
+## 7.9.4 — 2026-02-25
+
+### Patch changes
+
+- [a68a6a6](https://github.com/posthog/posthog-python/commit/a68a6a6d045072c88eeee7acac441536919b5954) feat(llma): add `$ai_tokens_source` property ("sdk" or "passthrough") to all `$ai_generation` events to detect when token values are externally overridden via `posthog_properties` — Thanks @carlos-marchal-ph!
+
+## 7.9.3 — 2026-02-18
+
+### Patch changes
+
+- [9f9553a](https://github.com/posthog/posthog-python/commit/9f9553a420d22e5e6435b775993f61a059280c2a) Fix posthoganalytics release, previously broken — Thanks @rafaeelaudibert!
+
+## 7.9.2 — 2026-02-18
+
+### Patch changes
+
+- [f1dc4d7](https://github.com/posthog/posthog-python/commit/f1dc4d73914712983a7f715ee4fe1b70e66e770a) Add sampo to the project — Thanks @rafaeelaudibert!
+
+## 7.9.1 - 2026-02-17
+
+fix(llma): make prompt fetches deterministic by requiring project_api_key and sending it as token query param
+
+## 7.9.0 - 2026-02-17
+
+feat: Support device_id as bucketing identifier for local evaluation
+
+## 7.8.6 - 2026-02-09
+
+fix: limit collections scanning in code variables
+
+## 7.8.5 - 2026-02-09
+
+fix: further optimize code variables pattern matching
+
+## 7.8.4 - 2026-02-09
+
+fix: do not pattern match long values in code variables
+
+## 7.8.3 - 2026-02-06
+
+fix: openAI input image sanitization
+
+## 7.8.2 - 2026-02-04
+
+fix(llma): fix prompts default url
+
+## 7.8.1 - 2026-02-03
+
+fix(llma): small fixes for prompt management
+
+## 7.8.0 - 2026-01-28
+
+feat(llma): add prompt management
+
+Adds the Prompt Management feature. At the time of release, this feature is in a closed alpha.
+
+## 7.7.0 - 2026-01-15
+
+feat(ai): Add OpenAI Agents SDK integration
+
+Automatic tracing for agent workflows, handoffs, tool calls, guardrails, and custom spans. Includes `$ai_total_tokens`, `$ai_error_type` categorization, and `$ai_framework` property.
+
+## 7.6.0 - 2026-01-12
+
+feat: add device_id to flags request payload
+
+Add device_id parameter to all feature flag methods, allowing the server to track device identifiers for flag evaluation. The device_id can be passed explicitly or set via context using `set_context_device_id()`.
+
+## 7.5.1 - 2026-01-07
+
+fix: avoid return from finally block to fix Python 3.14 SyntaxWarning (#361) - thanks @jodal
+
+## 7.5.0 - 2026-01-06
+
+feat: Capture Langchain, OpenAI and Anthropic errors as exceptions (if exception autocapture is enabled)
+feat: Add reference to exception in LLMA trace and span events
+
+## 7.4.3 - 2026-01-02
 
 Fixes cache creation cost for Langchain with Anthropic
 
-# 7.4.2 - 2025-12-22
+## 7.4.2 - 2025-12-22
 
 feat: add `in_app_modules` option to control code variables capturing
 
-# 7.4.1 - 2025-12-19
+## 7.4.1 - 2025-12-19
 
 fix: extract model from response for OpenAI stored prompts
 
 When using OpenAI stored prompts, the model is defined in the OpenAI dashboard rather than passed in the API request. This fix adds a fallback to extract the model from the response object when not provided in kwargs, ensuring generations show up with the correct model and enabling cost calculations.
 
-# 7.4.0 - 2025-12-16
+## 7.4.0 - 2025-12-16
 
 feat: Add automatic retries for feature flag requests
 
@@ -24,32 +122,32 @@ Feature flag API requests now automatically retry on transient failures:
 
 Rate limit (429) and quota (402) errors are not retried.
 
-# 7.3.1 - 2025-12-06
+## 7.3.1 - 2025-12-06
 
 fix: remove unused $exception_message and $exception_type
 
-# 7.3.0 - 2025-12-05
+## 7.3.0 - 2025-12-05
 
 feat: improve code variables capture masking
 
-# 7.2.0 - 2025-12-01
+## 7.2.0 - 2025-12-01
 
 feat: add $feature_flag_evaluated_at properties to $feature_flag_called events
 
-# 7.1.0 - 2025-11-26
+## 7.1.0 - 2025-11-26
 
 Add support for the async version of Gemini.
 
-# 7.0.2 - 2025-11-18
+## 7.0.2 - 2025-11-18
 
 Add support for Python 3.14.
 Projects upgrading to Python 3.14 should ensure any Pydantic models passed into the SDK use Pydantic v2, as Pydantic v1 is not compatible with Python 3.14.
 
-# 7.0.1 - 2025-11-15
+## 7.0.1 - 2025-11-15
 
 Try to use repr() when formatting code variables
 
-# 7.0.0 - 2025-11-11
+## 7.0.0 - 2025-11-11
 
 NB Python 3.9 is no longer supported
 
@@ -63,155 +161,155 @@ NB Python 3.9 is no longer supported
   - langchain-community: 0.3.29 → 0.4.1
   - langgraph: 0.6.6 → 1.0.2
 
-# 6.9.3 - 2025-11-10
+## 6.9.3 - 2025-11-10
 
 - feat(ph-ai): PostHog properties dict in GenerationMetadata
 
-# 6.9.2 - 2025-11-10
+## 6.9.2 - 2025-11-10
 
 - fix(llma): fix cache token double subtraction in Langchain for non-Anthropic providers causing negative costs
 
-# 6.9.1 - 2025-11-07
+## 6.9.1 - 2025-11-07
 
 - fix(error-tracking): pass code variables config from init to client
 
-# 6.9.0 - 2025-11-06
+## 6.9.0 - 2025-11-06
 
 - feat(error-tracking): add local variables capture
 
-# 6.8.0 - 2025-11-03
+## 6.8.0 - 2025-11-03
 
 - feat(llma): send web search calls to be used for LLM cost calculations
 
-# 6.7.14 - 2025-11-03
+## 6.7.14 - 2025-11-03
 
 - fix(django): Handle request.user access in async middleware context to prevent SynchronousOnlyOperation errors in Django 5+ (fixes #355)
 - test(django): Add Django 5 integration test suite with real ASGI application testing async middleware behavior
 
-# 6.7.13 - 2025-11-02
+## 6.7.13 - 2025-11-02
 
 - fix(llma): cache cost calculation in the LangChain callback
 
-# 6.7.12 - 2025-11-02
+## 6.7.12 - 2025-11-02
 
 - fix(django): Restore process_exception method to capture view and downstream middleware exceptions (fixes #329)
 - fix(ai/langchain): Add LangChain 1.0+ compatibility for CallbackHandler imports (fixes #362)
 
-# 6.7.11 - 2025-10-28
+## 6.7.11 - 2025-10-28
 
 - feat(ai): Add `$ai_framework` property for framework integrations (e.g. LangChain)
 
-# 6.7.10 - 2025-10-24
+## 6.7.10 - 2025-10-24
 
 - fix(django): Make middleware truly hybrid - compatible with both sync (WSGI) and async (ASGI) Django stacks without breaking sync-only deployments
 
-# 6.7.9 - 2025-10-22
+## 6.7.9 - 2025-10-22
 
 - fix(flags): multi-condition flags with static cohorts returning wrong variants
 
-# 6.7.8 - 2025-10-16
+## 6.7.8 - 2025-10-16
 
 - fix(llma): missing async for OpenAI's streaming implementation
 
-# 6.7.7 - 2025-10-14
+## 6.7.7 - 2025-10-14
 
 - fix: remove deprecated attribute $exception_personURL from exception events
 
-# 6.7.6 - 2025-09-16
+## 6.7.6 - 2025-09-16
 
 - fix: don't sort condition sets with variant overrides to the top
 - fix: Prevent core Client methods from raising exceptions
 
-# 6.7.5 - 2025-09-16
+## 6.7.5 - 2025-09-16
 
 - feat: Django middleware now supports async request handling.
 
-# 6.7.4 - 2025-09-05
+## 6.7.4 - 2025-09-05
 
 - fix: Missing system prompts for some providers
 
-# 6.7.3 - 2025-09-04
+## 6.7.3 - 2025-09-04
 
 - fix: missing usage tokens in Gemini
 
-# 6.7.2 - 2025-09-03
+## 6.7.2 - 2025-09-03
 
 - fix: tool call results in streaming providers
 
-# 6.7.1 - 2025-09-01
+## 6.7.1 - 2025-09-01
 
 - fix: Add base64 inline image sanitization
 
-# 6.7.0 - 2025-08-26
+## 6.7.0 - 2025-08-26
 
 - feat: Add support for feature flag dependencies
 
-# 6.6.1 - 2025-08-21
+## 6.6.1 - 2025-08-21
 
 - fix: Prevent `NoneType` error when `group_properties` is `None`
 
-# 6.6.0 - 2025-08-15
+## 6.6.0 - 2025-08-15
 
 - feat: Add `flag_keys_to_evaluate` parameter to optimize feature flag evaluation performance by only evaluating specified flags
 - feat: Add `flag_keys_filter` option to `send_feature_flags` for selective flag evaluation in capture events
 
-# 6.5.0 - 2025-08-08
+## 6.5.0 - 2025-08-08
 
 - feat: Add `$context_tags` to an event to know which properties were included as tags
 
-# 6.4.1 - 2025-08-06
+## 6.4.1 - 2025-08-06
 
 - fix: Always pass project API key in `remote_config` requests for deterministic project routing
 
-# 6.4.0 - 2025-08-05
+## 6.4.0 - 2025-08-05
 
 - feat: support Vertex AI for Gemini
 
-# 6.3.4 - 2025-08-04
+## 6.3.4 - 2025-08-04
 
 - fix: set `$ai_tools` for all providers and `$ai_output_choices` for all non-streaming provider flows properly
 
-# 6.3.3 - 2025-08-01
+## 6.3.3 - 2025-08-01
 
 - fix: `get_feature_flag_result` now correctly returns FeatureFlagResult when payload is empty string instead of None
 
-# 6.3.2 - 2025-07-31
+## 6.3.2 - 2025-07-31
 
 - fix: Anthropic's tool calls are now handled properly
 
-# 6.3.0 - 2025-07-22
+## 6.3.0 - 2025-07-22
 
 - feat: Enhanced `send_feature_flags` parameter to accept `SendFeatureFlagsOptions` object for declarative control over local/remote evaluation and custom properties
 
-# 6.2.1 - 2025-07-21
+## 6.2.1 - 2025-07-21
 
 - feat: make `posthog_client` an optional argument in PostHog AI providers wrappers (`posthog.ai.*`), intuitively using the default client as the default
 
-# 6.1.1 - 2025-07-16
+## 6.1.1 - 2025-07-16
 
 - fix: correctly capture exceptions processed by Django from views or middleware
 
-# 6.1.0 - 2025-07-10
+## 6.1.0 - 2025-07-10
 
 - feat: decouple feature flag local evaluation from personal API keys; support decrypting remote config payloads without relying on the feature flags poller
 
-# 6.0.4 - 2025-07-09
+## 6.0.4 - 2025-07-09
 
 - fix: add POSTHOG_MW_CLIENT setting to django middleware, to support custom clients for exception capture.
 
-# 6.0.3 - 2025-07-07
+## 6.0.3 - 2025-07-07
 
 - feat: add a feature flag evaluation cache (local storage or redis) to support returning flag evaluations when the service is down
 
-# 6.0.2 - 2025-07-02
+## 6.0.2 - 2025-07-02
 
 - fix: send_feature_flags changed to default to false in `Client::capture_exception`
 
-# 6.0.1
+## 6.0.1
 
 - fix: response `$process_person_profile` property when passed to capture
 
-# 6.0.0
+## 6.0.0
 
 This release contains a number of major breaking changes:
 
@@ -238,15 +336,15 @@ with posthog.new_context():
 
 Generally, arguments are now appropriately typed, and docstrings have been updated. If something is unclear, please open an issue, or submit a PR!
 
-# 5.4.0 - 2025-06-20
+## 5.4.0 - 2025-06-20
 
 - feat: add support to session_id context on page method
 
-# 5.3.0 - 2025-06-19
+## 5.3.0 - 2025-06-19
 
 - fix: safely handle exception values
 
-# 5.2.0 - 2025-06-19
+## 5.2.0 - 2025-06-19
 
 - feat: construct artificial stack traces if no traceback is available on a captured exception
 

CLAUDE.md

@@ -0,0 +1 @@
+LLM.md

LLM.md

@@ -0,0 +1,48 @@
+# Hanzo Insights Python SDK
+
+## Overview
+Integrate Hanzo Insights into any Python application. Package name: `hanzo-insights` on PyPI.
+
+## Tech Stack
+- **Language**: Python 3.10+
+- **Package**: `hanzo_insights` (import name), `hanzo-insights` (pip name)
+
+## Build & Run
+```bash
+uv sync
+uv run pytest
+```
+
+## Structure
+```
+posthog-python/
+  hanzo_insights/        # Main package
+    __init__.py          # Module-level API, Insights class
+    client.py            # Client class
+    ai/                  # AI provider integrations (OpenAI, Anthropic, Gemini, LangChain)
+    integrations/        # Framework integrations (Django middleware)
+    test/                # Tests
+  examples/
+  integration_tests/
+  pyproject.toml         # Package config (name: hanzo-insights)
+  setup.py               # Legacy setup
+```
+
+## Key Files
+- `pyproject.toml` -- Package config, dependencies, test config
+- `hanzo_insights/__init__.py` -- Public API surface
+- `hanzo_insights/client.py` -- Client implementation
+
+## Rebrand Notes
+- Main class: `Insights` (no backward compat aliases)
+- Django middleware: `InsightsContextMiddleware` (no backward compat aliases)
+- OpenAI Agents: `InsightsTracingProcessor` (no backward compat aliases)
+- `$lib` protocol value: `insights-python`
+- Ingestion URLs: `us.i.insights.hanzo.ai` / `eu.i.insights.hanzo.ai`
+- AI wrapper kwargs: `insights_*` (e.g. `insights_distinct_id`, `insights_trace_id`)
+- Exception attrs: `__insights_exception_captured`, `__insights_exception_uuid`
+- Context var: `insights_context_stack`
+- Redis prefix: `insights:flags:`
+- Redaction sentinels: `$$_insights_redacted_*`, `$$_insights_value_too_long_*`
+- Django settings: `INSIGHTS_MW_*` only (no `POSTHOG_MW_*` fallback)
+- Django headers: `X-INSIGHTS-SESSION-ID`, `X-INSIGHTS-DISTINCT-ID` only

Makefile

@@ -5,28 +5,42 @@ test:
 	coverage run -m pytest
 	coverage report
 
-release:
+build_release:
 	rm -rf dist/*
 	python setup.py sdist bdist_wheel
-	twine upload dist/*
 
-release_analytics:
+# Builds the `posthoganalytics` PyPI package, which is a mirror of `hanzo_insights`
+# published under a different name for backward compatibility with the upstream
+# posthog/posthog project.
+#
+# The process works in three phases:
+#   1. hanzo_insights -> posthoganalytics: Copy the source, rewrite all imports,
+#      remove the original hanzo_insights/ dir, and build the dist.
+#   2. posthoganalytics -> hanzo_insights: Reverse the import rewrites, copy
+#      everything back into hanzo_insights/, and clean up.
+#   3. Restore pyproject.toml from backup (setup_analytics.py modifies it).
+#
+# This ensures the working tree is left in the same state it started in.
+#
+# NOTE: This target clears dist/ before building. In the release workflow,
+# `build_release` (hanzo_insights) must be published BEFORE running this target,
+# otherwise the hanzo_insights dist artifacts will be lost.
+build_release_analytics:
 	rm -rf dist
 	rm -rf build
 	rm -rf posthoganalytics
 	mkdir posthoganalytics
-	cp -r posthog/* posthoganalytics/
-	find ./posthoganalytics -type f -name "*.py" -exec sed -i.bak -e 's/from posthog /from posthoganalytics /g' {} \;
-	find ./posthoganalytics -type f -name "*.py" -exec sed -i.bak -e 's/from posthog\./from posthoganalytics\./g' {} \;
+	cp -r hanzo_insights/* posthoganalytics/
+	find ./posthoganalytics -type f -name "*.py" -exec sed -i.bak -e 's/from hanzo_insights /from posthoganalytics /g' {} \;
+	find ./posthoganalytics -type f -name "*.py" -exec sed -i.bak -e 's/from hanzo_insights\./from posthoganalytics\./g' {} \;
 	find ./posthoganalytics -name "*.bak" -delete
-	rm -rf posthog
+	rm -rf hanzo_insights
 	python setup_analytics.py sdist bdist_wheel
-	twine upload dist/*
-	mkdir posthog
-	find ./posthoganalytics -type f -name "*.py" -exec sed -i.bak -e 's/from posthoganalytics /from posthog /g' {} \;
-	find ./posthoganalytics -type f -name "*.py" -exec sed -i.bak -e 's/from posthoganalytics\./from posthog\./g' {} \;
+	mkdir hanzo_insights
+	find ./posthoganalytics -type f -name "*.py" -exec sed -i.bak -e 's/from posthoganalytics /from hanzo_insights /g' {} \;
+	find ./posthoganalytics -type f -name "*.py" -exec sed -i.bak -e 's/from posthoganalytics\./from hanzo_insights\./g' {} \;
 	find ./posthoganalytics -name "*.bak" -delete
-	cp -r posthoganalytics/* posthog/
+	cp -r posthoganalytics/* hanzo_insights/
 	rm -rf posthoganalytics
 	rm -f pyproject.toml
 	cp pyproject.toml.backup pyproject.toml
@@ -41,17 +55,17 @@ prep_local:
 	cp -r . ../posthog-python-local/
 	cd ../posthog-python-local && rm -rf dist build posthoganalytics .git
 	cd ../posthog-python-local && mkdir posthoganalytics
-	cd ../posthog-python-local && cp -r posthog/* posthoganalytics/
-	cd ../posthog-python-local && find ./posthoganalytics -type f -name "*.py" -exec sed -i.bak -e 's/from posthog /from posthoganalytics /g' {} \;
-	cd ../posthog-python-local && find ./posthoganalytics -type f -name "*.py" -exec sed -i.bak -e 's/from posthog\./from posthoganalytics\./g' {} \;
+	cd ../posthog-python-local && cp -r hanzo_insights/* posthoganalytics/
+	cd ../posthog-python-local && find ./posthoganalytics -type f -name "*.py" -exec sed -i.bak -e 's/from hanzo_insights /from posthoganalytics /g' {} \;
+	cd ../posthog-python-local && find ./posthoganalytics -type f -name "*.py" -exec sed -i.bak -e 's/from hanzo_insights\./from posthoganalytics\./g' {} \;
 	cd ../posthog-python-local && find ./posthoganalytics -name "*.bak" -delete
-	cd ../posthog-python-local && rm -rf posthog
+	cd ../posthog-python-local && rm -rf hanzo_insights
 	cd ../posthog-python-local && sed -i.bak 's/from version import VERSION/from posthoganalytics.version import VERSION/' setup_analytics.py
 	cd ../posthog-python-local && rm setup_analytics.py.bak
-	cd ../posthog-python-local && sed -i.bak 's/"posthog"/"posthoganalytics"/' setup.py
+	cd ../posthog-python-local && sed -i.bak 's/"hanzo_insights"/"posthoganalytics"/' setup.py
 	cd ../posthog-python-local && rm setup.py.bak
 	cd ../posthog-python-local && python -c "import setup_analytics" 2>/dev/null || true
 	@echo "Local copy created at ../posthog-python-local"
 	@echo "Install with: pip install -e ../posthog-python-local"
 
-.PHONY: test lint release e2e_test prep_local
+.PHONY: test lint build_release build_release_analytics e2e_test prep_local

README.md

@@ -1,33 +1,53 @@
-# PostHog Python
+<p align="center"><img src=".github/hero.svg" alt="insights-python" width="880"></p>
 
-<p align="center">
-  <img alt="posthoglogo" src="https://user-images.githubusercontent.com/65415371/205059737-c8a4f836-4889-4654-902e-f302b187b6a0.png">
-</p>
-<p align="center">
-   <a href="https://pypi.org/project/posthog/"><img alt="pypi installs" src="https://img.shields.io/pypi/v/posthog"/></a>
-   <img alt="GitHub contributors" src="https://img.shields.io/github/contributors/posthog/posthog-python">
-  <img alt="GitHub commit activity" src="https://img.shields.io/github/commit-activity/m/posthog/posthog-python"/>
-  <img alt="GitHub closed issues" src="https://img.shields.io/github/issues-closed/posthog/posthog-python"/>
-</p>
+# Hanzo Insights Python SDK
 
-Please see the [Python integration docs](https://posthog.com/docs/integrations/python-integration) for details.
+Integrate [Hanzo Insights](https://insights.hanzo.ai) into any Python application.
+
+## Installation
+
+```bash
+pip install hanzo-insights
+```
+
+## Quick Start
+
+```python
+from hanzo_insights import Insights
+
+client = Insights('<your_project_api_key>', host='https://insights.hanzo.ai')
+
+# Capture an event
+client.capture('user_123', 'purchase', properties={'product': 'widget'})
+
+# Feature flags
+if client.feature_enabled('new-checkout', 'user_123'):
+    show_new_checkout()
+```
+
+## Module-level usage
+
+```python
+import hanzo_insights
+
+hanzo_insights.api_key = '<your_project_api_key>'
+hanzo_insights.host = 'https://insights.hanzo.ai'
+
+hanzo_insights.capture('movie_played', distinct_id='user_123', properties={'movie_id': '42'})
+hanzo_insights.shutdown()
+```
+
+## Python Version Support
+
+| SDK Version    | Python Versions Supported     |
+| -------------- | ----------------------------- |
+| 7.3.1+         | 3.10, 3.11, 3.12, 3.13, 3.14 |
+| 7.0.0 - 7.0.1  | 3.10, 3.11, 3.12, 3.13       |
+| 4.0.1 - 6.x    | 3.9, 3.10, 3.11, 3.12, 3.13  |
 
 ## Development
 
-### Testing Locally
-
-We recommend using [uv](https://docs.astral.sh/uv/). It's super fast.
-
-1. Run `uv venv env` (creates virtual environment called "env")
-    * or `python3 -m venv env`
-2. Run `source env/bin/activate` (activates the virtual environment)
-3. Run `uv sync --extra dev --extra test` (installs the package in develop mode, along with test dependencies)
-    * or `pip install -e ".[dev,test]"`
-4. you have to run `pre-commit install` to have auto linting pre commit
-5. Run `make test`
-  1. To run a specific test do `pytest -k test_no_api_key`
-
-## PostHog recommends `uv` so...
+We use [uv](https://docs.astral.sh/uv/).
 
 ```bash
 uv python install 3.12
@@ -39,28 +59,14 @@ pre-commit install
 make test
 ```
 
-### Running Locally
+### Running Tests
 
-Assuming you have a [local version of PostHog](https://posthog.com/docs/developing-locally) running, you can run `python3 example.py` to see the library in action.
-
-### Releasing Versions
-
-Updates are released automatically using GitHub Actions when `version.py` is updated on `master`. After bumping `version.py` in `master` and adding to `CHANGELOG.md`, the [release workflow](https://github.com/PostHog/posthog-python/blob/master/.github/workflows/release.yaml) will automatically trigger and deploy the new version.
-
-If you need to check the latest runs or manually trigger a release, you can go to [our release workflow's page](https://github.com/PostHog/posthog-python/actions/workflows/release.yaml) and dispatch it manually, using workflow from `master`.
-
-
-### Testing changes locally with the PostHog app
-
-You can run `make prep_local`, and it'll create a new folder alongside the SDK repo one called `posthog-python-local`, which you can then import into the posthog project by changing pyproject.toml to look like this:
-```toml
-dependencies = [
-    ...
-    "posthoganalytics" #NOTE: no version number
-    ...
-]
-...
-[tools.uv.sources]
-posthoganalytics = { path = "../posthog-python-local" }
+```bash
+make test
+# or run a specific test:
+pytest -k test_no_api_key
 ```
-This'll let you build and test SDK changes fully locally, incorporating them into your local posthog app stack. It mainly takes care of the `posthog -> posthoganalytics` module renaming. You'll need to re-run `make prep_local` each time you make a change, and re-run `uv sync --active` in the posthog app project.
+
+## License
+
+MIT

README_ANALYTICS.md

@@ -0,0 +1,11 @@
+# posthoganalytics
+
+> **Do not use this package.** Use [`posthog`](https://pypi.org/project/posthog/) instead.
+
+```bash
+pip install posthog
+```
+
+This package exists solely for internal use by [posthog/posthog](https://github.com/posthog/posthog) to avoid import conflicts with the local `posthog` package in that repository. It is an automatically generated mirror of `posthog` — same code, same versions, just published under a different name.
+
+If you are not working on the PostHog main repository, you should never need this package. All documentation, issues, and development happen in [`posthog-python`](https://github.com/posthog/posthog-python).

bin/build

@@ -4,5 +4,5 @@
 source bin/helpers/_utils.sh
 set_source_and_root_dir
 
-flake8 posthog --ignore E501,W503
+flake8 hanzo_insights --ignore E501,W503
 mypy --no-site-packages --config-file mypy.ini . | mypy-baseline filter

bin/docs

@@ -1,6 +1,6 @@
 #!/usr/bin/env bash
 #/ Usage: bin/docs
-#/ Description: Generate documentation for the PostHog Python SDK
+#/ Description: Generate documentation for the Insights Python SDK
 source bin/helpers/_utils.sh
 set_source_and_root_dir
 ensure_virtual_env

bin/docs_scripts/doc_constant.py

@@ -1,15 +1,15 @@
 """
-Constants for PostHog Python SDK documentation generation.
+Constants for Insights Python SDK documentation generation.
 """
 
 from typing import Dict, Union
-from posthog.version import VERSION
+from hanzo_insights.version import VERSION
 
 # Documentation generation metadata
 DOCUMENTATION_METADATA = {
     "hogRef": "0.3",
-    "slugPrefix": "posthog-python",
-    "specUrl": "https://github.com/PostHog/posthog-python",
+    "slugPrefix": "insights-python",
+    "specUrl": "https://github.com/Insights/insights-python",
 }
 
 # Docstring parsing patterns for new format
@@ -29,8 +29,8 @@ DOCSTRING_PATTERNS = {
 # Output file configuration
 OUTPUT_CONFIG: Dict[str, Union[str, int]] = {
     "output_dir": "./references",
-    "filename": f"posthog-python-references-{VERSION}.json",
-    "filename_latest": "posthog-python-references-latest.json",
+    "filename": f"insights-python-references-{VERSION}.json",
+    "filename_latest": "insights-python-references-latest.json",
     "indent": 2,
 }
 

bin/docs_scripts/generate_json_schemas.py

@@ -1,6 +1,6 @@
 #!/usr/bin/env python3
 """
-Generate comprehensive SDK documentation JSON from PostHog Python SDK.
+Generate comprehensive SDK documentation JSON from Insights Python SDK.
 This script inspects the code and docstrings to create documentation in the specified format.
 """
 
@@ -337,19 +337,19 @@ def analyze_type(cls) -> dict:
 def generate_sdk_documentation():
     """Generate complete SDK documentation in the requested format."""
 
-    # Import PostHog components
-    import posthog
-    from posthog.client import Client
-    import posthog.types as types_module
-    import posthog.args as args_module
-    from posthog.version import VERSION
+    # Import Insights components
+    import hanzo_insights
+    from hanzo_insights.client import Client
+    import hanzo_insights.types as types_module
+    import hanzo_insights.args as args_module
+    from hanzo_insights.version import VERSION
 
     # Main SDK info
     sdk_info = {
         "version": VERSION,
-        "id": "posthog-python",
-        "title": "PostHog Python SDK",
-        "description": "Integrate PostHog into any python application.",
+        "id": "insights-python",
+        "title": "Insights Python SDK",
+        "description": "Integrate Insights into any python application.",
         "slugPrefix": DOCUMENTATION_METADATA["slugPrefix"],
         "specUrl": DOCUMENTATION_METADATA["specUrl"],
     }
@@ -357,7 +357,7 @@ def generate_sdk_documentation():
     # Collect types
     types_list = []
 
-    # Types from posthog.types
+    # Types from hanzo_insights.types
     for name in dir(types_module):
         obj = getattr(types_module, name)
         if inspect.isclass(obj) and not name.startswith("_"):
@@ -367,7 +367,7 @@ def generate_sdk_documentation():
             except Exception as e:
                 print(f"Error analyzing type {name}: {e}")
 
-    # Types from posthog.args
+    # Types from hanzo_insights.args
     for name in dir(args_module):
         obj = getattr(args_module, name)
         if inspect.isclass(obj) and not name.startswith("_"):
@@ -388,26 +388,26 @@ def generate_sdk_documentation():
     # Collect classes
     classes_list = []
 
-    # Main PostHog class (renamed from Client)
+    # Main Insights class (renamed from Client)
     client_class = analyze_class(Client)
-    client_class["id"] = "PostHog"
-    client_class["title"] = "PostHog"
+    client_class["id"] = "Insights"
+    client_class["title"] = "Insights"
     classes_list.append(client_class)
 
-    # Global module functions (functions callable as posthog.function_name)
+    # Global module functions (functions callable as hanzo_insights.function_name)
     global_functions = []
-    for func_name in dir(posthog):
+    for func_name in dir(hanzo_insights):
         # Skip private functions and non-callables
-        if func_name.startswith("_") or not callable(getattr(posthog, func_name)):
+        if func_name.startswith("_") or not callable(getattr(hanzo_insights, func_name)):
             continue
 
-        func = getattr(posthog, func_name)
-        # Only include functions actually defined in the posthog module (not imported)
+        func = getattr(hanzo_insights, func_name)
+        # Only include functions actually defined in the hanzo_insights module (not imported)
         # and exclude class references
         if (
-            func_name not in ["Client", "Posthog"]
+            func_name not in ["Client", "Insights"]
             and hasattr(func, "__module__")
-            and func.__module__ == "posthog"
+            and func.__module__ == "hanzo_insights"
         ):
             try:
                 func_info = analyze_function(func, func_name)
@@ -421,8 +421,8 @@ def generate_sdk_documentation():
         classes_list.append(
             {
                 "id": "PostHogModule",
-                "title": "PostHog Module Functions",
-                "description": "Global functions available in the PostHog module",
+                "title": "Insights Module Functions",
+                "description": "Global functions available in the Insights module",
                 "functions": global_functions,
             }
         )
@@ -443,7 +443,7 @@ def generate_sdk_documentation():
 
     # Create the final structure
     result = {
-        "id": "posthog-python",
+        "id": "insights-python",
         "hogRef": DOCUMENTATION_METADATA["hogRef"],
         "info": sdk_info,
         "types": types_list,
@@ -455,7 +455,7 @@ def generate_sdk_documentation():
 
 
 if __name__ == "__main__":
-    print("Generating PostHog Python SDK documentation...")
+    print("Generating Insights Python SDK documentation...")
 
     try:
         documentation = generate_sdk_documentation()

example.py

@@ -1,18 +1,18 @@
-# PostHog Python library example
+# Hanzo Insights Python library example
 #
-# This script demonstrates various PostHog Python SDK capabilities including:
+# This script demonstrates various Hanzo Insights Python SDK capabilities including:
 # - Basic event capture and user identification
 # - Feature flag local evaluation
 # - Feature flag payloads
 # - Context management and tagging
 #
 # Setup:
-# 1. Copy .env.example to .env and fill in your PostHog credentials
+# 1. Copy .env.example to .env and fill in your Insights credentials
 # 2. Run this script and choose from the interactive menu
 
 import os
 
-import posthog
+import hanzo_insights
 
 
 def load_env_file():
@@ -31,30 +31,30 @@ def load_env_file():
 load_env_file()
 
 # Get configuration
-project_key = os.getenv("POSTHOG_PROJECT_API_KEY", "")
-personal_api_key = os.getenv("POSTHOG_PERSONAL_API_KEY", "")
-host = os.getenv("POSTHOG_HOST", "http://localhost:8000")
+project_key = os.getenv("INSIGHTS_PROJECT_API_KEY", "")
+personal_api_key = os.getenv("INSIGHTS_PERSONAL_API_KEY", "")
+host = os.getenv("INSIGHTS_HOST", "http://localhost:8000")
 
 # Check if project key is provided (required)
 if not project_key:
-    print("❌ Missing PostHog project API key!")
-    print("   Please set POSTHOG_PROJECT_API_KEY environment variable")
+    print("❌ Missing Insights project API key!")
+    print("   Please set INSIGHTS_PROJECT_API_KEY environment variable")
     print("   or copy .env.example to .env and fill in your values")
     exit(1)
 
-# Configure PostHog with credentials
-posthog.debug = False
-posthog.api_key = project_key
-posthog.project_api_key = project_key
-posthog.host = host
-posthog.poll_interval = 10
+# Configure Insights with credentials
+hanzo_insights.debug = False
+hanzo_insights.api_key = project_key
+hanzo_insights.project_api_key = project_key
+hanzo_insights.host = host
+hanzo_insights.poll_interval = 10
 
 # Check if personal API key is available for local evaluation
 local_eval_available = bool(personal_api_key)
 if personal_api_key:
-    posthog.personal_api_key = personal_api_key
+    hanzo_insights.personal_api_key = personal_api_key
 
-print("🔑 PostHog Configuration:")
+print("🔑 Insights Configuration:")
 print(f"   Project API Key: {project_key[:9]}...")
 if local_eval_available:
     print("   Personal API Key: [SET]")
@@ -63,7 +63,7 @@ else:
 print(f"   Host: {host}\n")
 
 # Display menu and get user choice
-print("🚀 PostHog Python SDK Demo - Choose an example to run:\n")
+print("🚀 Hanzo Insights Python SDK Demo - Choose an example to run:\n")
 print("1. Identify and capture examples")
 local_eval_note = "" if local_eval_available else " [requires personal API key]"
 print(f"2. Feature flag local evaluation examples{local_eval_note}")
@@ -79,11 +79,11 @@ if choice == "1":
     print("IDENTIFY AND CAPTURE EXAMPLES")
     print("=" * 60)
 
-    posthog.debug = True
+    hanzo_insights.debug = True
 
     # Capture an event
     print("📊 Capturing events...")
-    posthog.capture(
+    hanzo_insights.capture(
         "event",
         distinct_id="distinct_id",
         properties={"property1": "value", "property2": "value"},
@@ -92,14 +92,14 @@ if choice == "1":
 
     # Alias a previous distinct id with a new one
     print("🔗 Creating alias...")
-    posthog.alias("distinct_id", "new_distinct_id")
+    hanzo_insights.alias("distinct_id", "new_distinct_id")
 
-    posthog.capture(
+    hanzo_insights.capture(
         "event2",
         distinct_id="new_distinct_id",
         properties={"property1": "value", "property2": "value"},
     )
-    posthog.capture(
+    hanzo_insights.capture(
         "event-with-groups",
         distinct_id="new_distinct_id",
         properties={"property1": "value", "property2": "value"},
@@ -108,28 +108,28 @@ if choice == "1":
 
     # Add properties to the person
     print("👤 Identifying user...")
-    posthog.set(
+    hanzo_insights.set(
         distinct_id="new_distinct_id", properties={"email": "something@something.com"}
     )
 
     # Add properties to a group
     print("🏢 Identifying group...")
-    posthog.group_identify("company", "id:5", {"employees": 11})
+    hanzo_insights.group_identify("company", "id:5", {"employees": 11})
 
     # Properties set only once to the person
     print("🔒 Setting properties once...")
-    posthog.set_once(
+    hanzo_insights.set_once(
         distinct_id="new_distinct_id", properties={"self_serve_signup": True}
     )
 
     # This will not change the property (because it was already set)
-    posthog.set_once(
+    hanzo_insights.set_once(
         distinct_id="new_distinct_id", properties={"self_serve_signup": False}
     )
 
     print("🔄 Updating properties...")
-    posthog.set(distinct_id="new_distinct_id", properties={"current_browser": "Chrome"})
-    posthog.set(
+    hanzo_insights.set(distinct_id="new_distinct_id", properties={"current_browser": "Chrome"})
+    hanzo_insights.set(
         distinct_id="new_distinct_id", properties={"current_browser": "Firefox"}
     )
 
@@ -137,45 +137,45 @@ elif choice == "2":
     if not local_eval_available:
         print("\n❌ This example requires a personal API key for local evaluation.")
         print(
-            "   Set POSTHOG_PERSONAL_API_KEY environment variable to run this example."
+            "   Set INSIGHTS_PERSONAL_API_KEY environment variable to run this example."
         )
-        posthog.shutdown()
+        hanzo_insights.shutdown()
         exit(1)
 
     print("\n" + "=" * 60)
     print("FEATURE FLAG LOCAL EVALUATION EXAMPLES")
     print("=" * 60)
 
-    posthog.debug = True
+    hanzo_insights.debug = True
 
     print("🏁 Testing basic feature flags...")
     print(
-        f"beta-feature for 'distinct_id': {posthog.feature_enabled('beta-feature', 'distinct_id')}"
+        f"beta-feature for 'distinct_id': {hanzo_insights.feature_enabled('beta-feature', 'distinct_id')}"
     )
     print(
-        f"beta-feature for 'new_distinct_id': {posthog.feature_enabled('beta-feature', 'new_distinct_id')}"
+        f"beta-feature for 'new_distinct_id': {hanzo_insights.feature_enabled('beta-feature', 'new_distinct_id')}"
     )
     print(
-        f"beta-feature with groups: {posthog.feature_enabled('beta-feature-groups', 'distinct_id', groups={'company': 'id:5'})}"
+        f"beta-feature with groups: {hanzo_insights.feature_enabled('beta-feature-groups', 'distinct_id', groups={'company': 'id:5'})}"
     )
 
     print("\n🌍 Testing location-based flags...")
     # Assume test-flag has `City Name = Sydney` as a person property set
     print(
-        f"Sydney user: {posthog.feature_enabled('test-flag', 'random_id_12345', person_properties={'$geoip_city_name': 'Sydney'})}"
+        f"Sydney user: {hanzo_insights.feature_enabled('test-flag', 'random_id_12345', person_properties={'$geoip_city_name': 'Sydney'})}"
     )
 
     print(
-        f"Sydney user (local only): {posthog.feature_enabled('test-flag', 'distinct_id_random_22', person_properties={'$geoip_city_name': 'Sydney'}, only_evaluate_locally=True)}"
+        f"Sydney user (local only): {hanzo_insights.feature_enabled('test-flag', 'distinct_id_random_22', person_properties={'$geoip_city_name': 'Sydney'}, only_evaluate_locally=True)}"
     )
 
     print("\n📋 Getting all flags...")
-    print(f"All flags: {posthog.get_all_flags('distinct_id_random_22')}")
+    print(f"All flags: {hanzo_insights.get_all_flags('distinct_id_random_22')}")
     print(
-        f"All flags (local): {posthog.get_all_flags('distinct_id_random_22', only_evaluate_locally=True)}"
+        f"All flags (local): {hanzo_insights.get_all_flags('distinct_id_random_22', only_evaluate_locally=True)}"
     )
     print(
-        f"All flags with properties: {posthog.get_all_flags('distinct_id_random_22', person_properties={'$geoip_city_name': 'Sydney'}, only_evaluate_locally=True)}"
+        f"All flags with properties: {hanzo_insights.get_all_flags('distinct_id_random_22', person_properties={'$geoip_city_name': 'Sydney'}, only_evaluate_locally=True)}"
     )
 
 elif choice == "3":
@@ -183,22 +183,22 @@ elif choice == "3":
     print("FEATURE FLAG PAYLOAD EXAMPLES")
     print("=" * 60)
 
-    posthog.debug = True
+    hanzo_insights.debug = True
 
     print("📦 Testing feature flag payloads...")
     print(
-        f"beta-feature payload: {posthog.get_feature_flag_payload('beta-feature', 'distinct_id')}"
+        f"beta-feature payload: {hanzo_insights.get_feature_flag_payload('beta-feature', 'distinct_id')}"
     )
     print(
-        f"All flags and payloads: {posthog.get_all_flags_and_payloads('distinct_id')}"
+        f"All flags and payloads: {hanzo_insights.get_all_flags_and_payloads('distinct_id')}"
     )
     print(
-        f"Remote config payload: {posthog.get_remote_config_payload('encrypted_payload_flag_key')}"
+        f"Remote config payload: {hanzo_insights.get_remote_config_payload('encrypted_payload_flag_key')}"
     )
 
     # Get feature flag result with all details (enabled, variant, payload, key, reason)
     print("\n🔍 Getting detailed flag result...")
-    result = posthog.get_feature_flag_result("beta-feature", "distinct_id")
+    result = hanzo_insights.get_feature_flag_result("beta-feature", "distinct_id")
     if result:
         print(f"Flag key: {result.key}")
         print(f"Flag enabled: {result.enabled}")
@@ -212,9 +212,9 @@ elif choice == "4":
     if not local_eval_available:
         print("\n❌ This example requires a personal API key for local evaluation.")
         print(
-            "   Set POSTHOG_PERSONAL_API_KEY environment variable to run this example."
+            "   Set INSIGHTS_PERSONAL_API_KEY environment variable to run this example."
         )
-        posthog.shutdown()
+        hanzo_insights.shutdown()
         exit(1)
 
     print("\n" + "=" * 60)
@@ -234,10 +234,10 @@ elif choice == "4":
     print("      - Rollout: 100%")
     print("")
 
-    posthog.debug = True
+    hanzo_insights.debug = True
 
     # Test @example.com user (should satisfy dependency if flags exist)
-    result1 = posthog.feature_enabled(
+    result1 = hanzo_insights.feature_enabled(
         "test-flag-dependency",
         "example_user",
         person_properties={"email": "user@example.com"},
@@ -246,7 +246,7 @@ elif choice == "4":
     print(f"✅ @example.com user (test-flag-dependency): {result1}")
 
     # Test non-example.com user (dependency should not be satisfied)
-    result2 = posthog.feature_enabled(
+    result2 = hanzo_insights.feature_enabled(
         "test-flag-dependency",
         "regular_user",
         person_properties={"email": "user@other.com"},
@@ -255,13 +255,13 @@ elif choice == "4":
     print(f"❌ Regular user (test-flag-dependency): {result2}")
 
     # Test beta-feature directly for comparison
-    beta1 = posthog.feature_enabled(
+    beta1 = hanzo_insights.feature_enabled(
         "beta-feature",
         "example_user",
         person_properties={"email": "user@example.com"},
         only_evaluate_locally=True,
     )
-    beta2 = posthog.feature_enabled(
+    beta2 = hanzo_insights.feature_enabled(
         "beta-feature",
         "regular_user",
         person_properties={"email": "user@other.com"},
@@ -303,7 +303,7 @@ elif choice == "4":
     print("")
 
     # Test pineapple -> blue -> breaking-bad chain
-    dependent_result3 = posthog.get_feature_flag(
+    dependent_result3 = hanzo_insights.get_feature_flag(
         "multivariate-root-flag",
         "regular_user",
         person_properties={"email": "pineapple@example.com"},
@@ -317,7 +317,7 @@ elif choice == "4":
         print("✅ 'multivariate-root-flag' with email pineapple@example.com succeeded")
 
     # Test mango -> red -> the-wire chain
-    dependent_result4 = posthog.get_feature_flag(
+    dependent_result4 = hanzo_insights.get_feature_flag(
         "multivariate-root-flag",
         "regular_user",
         person_properties={"email": "mango@example.com"},
@@ -336,19 +336,19 @@ elif choice == "4":
         ("pineapple@example.com", ["pineapple", "blue", "breaking-bad"]),
         ("mango@example.com", ["mango", "red", "the-wire"]),
     ]:
-        leaf = posthog.get_feature_flag(
+        leaf = hanzo_insights.get_feature_flag(
             "multivariate-leaf-flag",
             "regular_user",
             person_properties={"email": email},
             only_evaluate_locally=True,
         )
-        intermediate = posthog.get_feature_flag(
+        intermediate = hanzo_insights.get_feature_flag(
             "multivariate-intermediate-flag",
             "regular_user",
             person_properties={"email": email},
             only_evaluate_locally=True,
         )
-        root = posthog.get_feature_flag(
+        root = hanzo_insights.get_feature_flag(
             "multivariate-root-flag",
             "regular_user",
             person_properties={"email": email},
@@ -373,7 +373,7 @@ elif choice == "5":
     print("CONTEXT MANAGEMENT AND TAGGING EXAMPLES")
     print("=" * 60)
 
-    posthog.debug = True
+    hanzo_insights.debug = True
 
     print("🏷️ Testing context management...")
     print(
@@ -384,12 +384,12 @@ elif choice == "5":
     # and tagged with the context tags. Other events captured will also be tagged with the context tags. By default,
     # the new context inherits tags from the parent context.
     try:
-        with posthog.new_context():
-            posthog.tag("transaction_id", "abc123")
-            posthog.tag("some_arbitrary_value", {"tags": "can be dicts"})
+        with hanzo_insights.new_context():
+            hanzo_insights.tag("transaction_id", "abc123")
+            hanzo_insights.tag("some_arbitrary_value", {"tags": "can be dicts"})
 
             # This event will be captured with the tags set above
-            posthog.capture("order_processed")
+            hanzo_insights.capture("order_processed")
             print("✅ Event captured with inherited context tags")
             # This exception will be captured with the tags set above
             # raise Exception("Order processing failed")
@@ -398,30 +398,30 @@ elif choice == "5":
 
     # Use fresh=True to start with a clean context (no inherited tags)
     try:
-        with posthog.new_context(fresh=True):
-            posthog.tag("session_id", "xyz789")
+        with hanzo_insights.new_context(fresh=True):
+            hanzo_insights.tag("session_id", "xyz789")
             # Only session_id tag will be present, no inherited tags
-            posthog.capture("session_event")
+            hanzo_insights.capture("session_event")
             print("✅ Event captured with fresh context tags")
             # raise Exception("Session handling failed")
     except Exception as e:
         print(f"Exception captured: {e}")
 
-    # You can also use the `@posthog.scoped()` decorator to enter a new context.
+    # You can also use the `@hanzo_insights.scoped()` decorator to enter a new context.
     # By default, it inherits tags from the parent context
-    @posthog.scoped()
+    @hanzo_insights.scoped()
     def process_order(order_id):
-        posthog.tag("order_id", order_id)
-        posthog.capture("order_step_completed")
+        hanzo_insights.tag("order_id", order_id)
+        hanzo_insights.capture("order_step_completed")
         print(f"✅ Order {order_id} processed with scoped context")
         # Exception will be captured and tagged automatically
         # raise Exception("Order processing failed")
 
     # Use fresh=True to start with a clean context (no inherited tags)
-    @posthog.scoped(fresh=True)
+    @hanzo_insights.scoped(fresh=True)
     def process_payment(payment_id):
-        posthog.tag("payment_id", payment_id)
-        posthog.capture("payment_processed")
+        hanzo_insights.tag("payment_id", payment_id)
+        hanzo_insights.capture("payment_processed")
         print(f"✅ Payment {payment_id} processed with fresh scoped context")
         # Only payment_id tag will be present, no inherited tags
         # raise Exception("Payment processing failed")
@@ -436,18 +436,18 @@ elif choice == "6":
 
     # Run example 1
     print(f"\n{'🔸' * 20} IDENTIFY AND CAPTURE {'🔸' * 20}")
-    posthog.debug = True
+    hanzo_insights.debug = True
     print("📊 Capturing events...")
-    posthog.capture(
+    hanzo_insights.capture(
         "event",
         distinct_id="distinct_id",
         properties={"property1": "value", "property2": "value"},
         send_feature_flags=True,
     )
     print("🔗 Creating alias...")
-    posthog.alias("distinct_id", "new_distinct_id")
+    hanzo_insights.alias("distinct_id", "new_distinct_id")
     print("👤 Identifying user...")
-    posthog.set(
+    hanzo_insights.set(
         distinct_id="new_distinct_id", properties={"email": "something@something.com"}
     )
 
@@ -455,27 +455,27 @@ elif choice == "6":
     if local_eval_available:
         print(f"\n{'🔸' * 20} FEATURE FLAGS {'🔸' * 20}")
         print("🏁 Testing basic feature flags...")
-        print(f"beta-feature: {posthog.feature_enabled('beta-feature', 'distinct_id')}")
+        print(f"beta-feature: {hanzo_insights.feature_enabled('beta-feature', 'distinct_id')}")
         print(
-            f"Sydney user: {posthog.feature_enabled('test-flag', 'random_id_12345', person_properties={'$geoip_city_name': 'Sydney'})}"
+            f"Sydney user: {hanzo_insights.feature_enabled('test-flag', 'random_id_12345', person_properties={'$geoip_city_name': 'Sydney'})}"
         )
 
     # Run example 3
     print(f"\n{'🔸' * 20} PAYLOADS {'🔸' * 20}")
     print("📦 Testing payloads...")
-    print(f"Payload: {posthog.get_feature_flag_payload('beta-feature', 'distinct_id')}")
+    print(f"Payload: {hanzo_insights.get_feature_flag_payload('beta-feature', 'distinct_id')}")
 
     # Run example 4 (requires local evaluation)
     if local_eval_available:
         print(f"\n{'🔸' * 20} FLAG DEPENDENCIES {'🔸' * 20}")
         print("🔗 Testing flag dependencies...")
-        result1 = posthog.feature_enabled(
+        result1 = hanzo_insights.feature_enabled(
             "test-flag-dependency",
             "demo_user",
             person_properties={"email": "user@example.com"},
             only_evaluate_locally=True,
         )
-        result2 = posthog.feature_enabled(
+        result2 = hanzo_insights.feature_enabled(
             "test-flag-dependency",
             "demo_user2",
             person_properties={"email": "user@other.com"},
@@ -486,23 +486,23 @@ elif choice == "6":
     # Run example 5
     print(f"\n{'🔸' * 20} CONTEXT MANAGEMENT {'🔸' * 20}")
     print("🏷️ Testing context management...")
-    with posthog.new_context():
-        posthog.tag("demo_run", "all_examples")
-        posthog.capture("demo_completed")
+    with hanzo_insights.new_context():
+        hanzo_insights.tag("demo_run", "all_examples")
+        hanzo_insights.capture("demo_completed")
         print("✅ Demo completed with context tags")
 
 elif choice == "7":
     print("👋 Goodbye!")
-    posthog.shutdown()
+    hanzo_insights.shutdown()
     exit()
 
 else:
     print("❌ Invalid choice. Please run again and select 1-7.")
-    posthog.shutdown()
+    hanzo_insights.shutdown()
     exit()
 
 print("\n" + "=" * 60)
 print("✅ Example completed!")
 print("=" * 60)
 
-posthog.shutdown()
+hanzo_insights.shutdown()

examples/redis_flag_cache.py

@@ -1,17 +1,17 @@
 """
-Redis-based distributed cache for PostHog feature flag definitions.
+Redis-based distributed cache for Insights feature flag definitions.
 
 This example demonstrates how to implement a FlagDefinitionCacheProvider
 using Redis for multi-instance deployments (leader election pattern).
 
 Usage:
     import redis
-    from posthog import Posthog
+    from hanzo_insights import Insights
 
     redis_client = redis.Redis(host='localhost', port=6379, decode_responses=True)
     cache = RedisFlagCache(redis_client, service_key="my-service")
 
-    posthog = Posthog(
+    client = Insights(
         "<project_api_key>",
         personal_api_key="<personal_api_key>",
         flag_definition_cache_provider=cache,
@@ -24,17 +24,17 @@ Requirements:
 import json
 import uuid
 
-from posthog import FlagDefinitionCacheData, FlagDefinitionCacheProvider
+from hanzo_insights import FlagDefinitionCacheData, FlagDefinitionCacheProvider
 from redis import Redis
 from typing import Optional
 
 
 class RedisFlagCache(FlagDefinitionCacheProvider):
     """
-    A distributed cache for PostHog feature flag definitions using Redis.
+    A distributed cache for Insights feature flag definitions using Redis.
 
     In a multi-instance deployment (e.g., multiple serverless functions or containers),
-    we want only ONE instance to poll PostHog for flag updates, while all instances
+    we want only ONE instance to poll Insights for flag updates, while all instances
     share the cached results. This prevents N instances from making N redundant API calls.
 
     The implementation uses leader election:
@@ -83,8 +83,8 @@ class RedisFlagCache(FlagDefinitionCacheProvider):
                          Examples: "my-api-prod", "checkout-service", "staging".
 
         Redis Keys Created:
-            - posthog:flags:{service_key} - Cached flag definitions (JSON)
-            - posthog:flags:{service_key}:lock - Leader election lock
+            - insights:flags:{service_key} - Cached flag definitions (JSON)
+            - insights:flags:{service_key}:lock - Leader election lock
 
         Example:
             redis_client = redis.Redis(
@@ -95,8 +95,8 @@ class RedisFlagCache(FlagDefinitionCacheProvider):
             cache = RedisFlagCache(redis_client, service_key="my-api-prod")
         """
         self._redis = redis
-        self._cache_key = f"posthog:flags:{service_key}"
-        self._lock_key = f"posthog:flags:{service_key}:lock"
+        self._cache_key = f"insights:flags:{service_key}"
+        self._lock_key = f"insights:flags:{service_key}:lock"
         self._instance_id = str(uuid.uuid4())
         self._try_lead = self._redis.register_script(self._LUA_TRY_LEAD)
         self._stop_lead = self._redis.register_script(self._LUA_STOP_LEAD)
@@ -113,7 +113,7 @@ class RedisFlagCache(FlagDefinitionCacheProvider):
 
     def should_fetch_flag_definitions(self) -> bool:
         """
-        Determines if this instance should fetch flag definitions from PostHog.
+        Determines if this instance should fetch flag definitions from Insights.
 
         Atomically either:
         - Acquires the lock if no one holds it, OR

examples/remote_config.py

@@ -1,15 +1,15 @@
 #!/usr/bin/env python3
 """
-Simple test script for PostHog remote config endpoint.
+Simple test script for Insights remote config endpoint.
 """
 
-import posthog
+import hanzo_insights
 
-# Initialize PostHog client
-posthog.api_key = "phc_..."
-posthog.personal_api_key = "phs_..."  # or "phx_..."
-posthog.host = "http://localhost:8000"  # or "https://us.posthog.com"
-posthog.debug = True
+# Initialize Insights client
+hanzo_insights.api_key = "phc_..."
+hanzo_insights.personal_api_key = "phs_..."  # or "phx_..."
+hanzo_insights.host = "http://localhost:8000"  # or "https://us.insights.hanzo.ai"
+hanzo_insights.debug = True
 
 
 def test_remote_config():
@@ -21,7 +21,7 @@ def test_remote_config():
 
     try:
         # Get remote config payload
-        payload = posthog.get_remote_config_payload(flag_key)
+        payload = hanzo_insights.get_remote_config_payload(flag_key)
         print(f"✅ Success! Remote config payload for '{flag_key}': {payload}")
 
     except Exception as e:

hanzo_insights/__init__.py

@@ -3,60 +3,66 @@ from typing import Any, Callable, Dict, Optional  # noqa: F401
 
 from typing_extensions import Unpack
 
-from posthog.args import ExceptionArg, OptionalCaptureArgs, OptionalSetArgs
-from posthog.client import Client
-from posthog.contexts import (
+from hanzo_insights.args import ExceptionArg, OptionalCaptureArgs, OptionalSetArgs
+from hanzo_insights.client import Client
+from hanzo_insights.contexts import (
     identify_context as inner_identify_context,
 )
-from posthog.contexts import (
+from hanzo_insights.contexts import (
     new_context as inner_new_context,
 )
-from posthog.contexts import (
+from hanzo_insights.contexts import (
     scoped as inner_scoped,
 )
-from posthog.contexts import (
+from hanzo_insights.contexts import (
     set_capture_exception_code_variables_context as inner_set_capture_exception_code_variables_context,
 )
-from posthog.contexts import (
+from hanzo_insights.contexts import (
     set_code_variables_ignore_patterns_context as inner_set_code_variables_ignore_patterns_context,
 )
-from posthog.contexts import (
+from hanzo_insights.contexts import (
     set_code_variables_mask_patterns_context as inner_set_code_variables_mask_patterns_context,
 )
-from posthog.contexts import (
+from hanzo_insights.contexts import (
+    set_context_device_id as inner_set_context_device_id,
+)
+from hanzo_insights.contexts import (
     set_context_session as inner_set_context_session,
 )
-from posthog.contexts import (
+from hanzo_insights.contexts import (
     tag as inner_tag,
 )
-from posthog.exception_utils import (
+from hanzo_insights.contexts import (
+    get_tags as inner_get_tags,
+)
+from hanzo_insights.exception_utils import (
     DEFAULT_CODE_VARIABLES_IGNORE_PATTERNS,
     DEFAULT_CODE_VARIABLES_MASK_PATTERNS,
 )
-from posthog.feature_flags import (
+from hanzo_insights.feature_flags import (
     InconclusiveMatchError as InconclusiveMatchError,
 )
-from posthog.feature_flags import (
+from hanzo_insights.feature_flags import (
     RequiresServerEvaluation as RequiresServerEvaluation,
 )
-from posthog.flag_definition_cache import (
+from hanzo_insights.flag_definition_cache import (
     FlagDefinitionCacheData as FlagDefinitionCacheData,
     FlagDefinitionCacheProvider as FlagDefinitionCacheProvider,
 )
-from posthog.request import (
+from hanzo_insights.request import (
     disable_connection_reuse as disable_connection_reuse,
     enable_keep_alive as enable_keep_alive,
     set_socket_options as set_socket_options,
     SocketOptions as SocketOptions,
 )
-from posthog.types import (
+from hanzo_insights.types import (
     FeatureFlag,
     FlagsAndPayloads,
 )
-from posthog.types import (
+from hanzo_insights.types import (
     FeatureFlagResult as FeatureFlagResult,
 )
-from posthog.version import VERSION
+from hanzo_insights.version import VERSION
 
 __version__ = VERSION
 
@@ -70,11 +76,11 @@ def new_context(fresh=False, capture_exceptions=True, client=None):
     Args:
         fresh: Whether to start with a fresh context (default: False)
         capture_exceptions: Whether to capture exceptions raised within the context (default: True)
-        client: Optional Posthog client instance to use for this context (default: None)
+        client: Optional Insights client instance to use for this context (default: None)
 
     Examples:
         ```python
-        from posthog import new_context, tag, capture
+        from hanzo_insights import new_context, tag, capture
         with new_context():
             tag("request_id", "123")
             capture("event_name", properties={"property": "value"})
@@ -94,11 +100,11 @@ def scoped(fresh=False, capture_exceptions=True):
 
     Args:
         fresh: Whether to start with a fresh context (default: False)
-        capture_exceptions: Whether to capture and track exceptions with posthog error tracking (default: True)
+        capture_exceptions: Whether to capture and track exceptions with Insights error tracking (default: True)
 
     Examples:
         ```python
-        from posthog import scoped, tag, capture
+        from hanzo_insights import scoped, tag, capture
         @scoped()
         def process_payment(payment_id):
             tag("payment_id", payment_id)
@@ -120,7 +126,7 @@ def set_context_session(session_id: str):
 
     Examples:
         ```python
-        from posthog import set_context_session
+        from hanzo_insights import set_context_session
         set_context_session("session_123")
         ```
 
@@ -130,6 +136,26 @@ def set_context_session(session_id: str):
     return inner_set_context_session(session_id)
 
 
+def set_context_device_id(device_id: str):
+    """
+    Set the device ID for the current context, associating all feature flag requests
+    in this or child contexts with the given device ID.
+
+    Args:
+        device_id: The device ID to associate with the current context and its children
+
+    Examples:
+        ```python
+        from hanzo_insights import set_context_device_id
+        set_context_device_id("device_123")
+        ```
+
+    Category:
+        Contexts
+    """
+    return inner_set_context_device_id(device_id)
+
+
 def identify_context(distinct_id: str):
     """
     Identify the current context with a distinct ID.
@@ -139,7 +165,7 @@ def identify_context(distinct_id: str):
 
     Examples:
         ```python
-        from posthog import identify_context
+        from hanzo_insights import identify_context
         identify_context("user_123")
         ```
 
@@ -180,7 +206,7 @@ def tag(name: str, value: Any):
 
     Examples:
         ```python
-        from posthog import tag
+        from hanzo_insights import tag
         tag("user_id", "123")
         ```
 
@@ -190,6 +216,19 @@ def tag(name: str, value: Any):
     return inner_tag(name, value)
 
 
+def get_tags() -> Dict[str, Any]:
+    """
+    Get all tags from the current context.
+
+    Returns:
+        Dict of all tags in the current context
+
+    Category:
+        Contexts
+    """
+    return inner_get_tags()
+
+
 """Settings."""
 api_key = None  # type: Optional[str]
 host = None  # type: Optional[str]
@@ -224,7 +263,7 @@ in_app_modules = None  # type: Optional[list[str]]
 
 
 # NOTE - this and following functions take unpacked kwargs because we needed to make
-# it impossible to write `posthog.capture(distinct-id, event-name)` - basically, to enforce
+# it impossible to write `hanzo_insights.capture(distinct-id, event-name)` - basically, to enforce
 # the breaking change made between 5.3.0 and 6.0.0. This decision can be unrolled in later
 # versions, without a breaking change, to get back the type information in function signatures
 def capture(event: str, **kwargs: Unpack[OptionalCaptureArgs]) -> Optional[str]:
@@ -241,12 +280,12 @@ def capture(event: str, **kwargs: Unpack[OptionalCaptureArgs]) -> Optional[str]:
             disable_geoip: Whether to disable GeoIP lookup
 
     Details:
-        Capture allows you to capture anything a user does within your system, which you can later use in PostHog to find patterns in usage, work out which features to improve or where people are giving up. A capture call requires an event name to specify the event. We recommend using [verb] [noun], like `movie played` or `movie updated` to easily identify what your events mean later on. Capture takes a number of optional arguments, which are defined by the `OptionalCaptureArgs` type.
+        Capture allows you to capture anything a user does within your system, which you can later use in Insights to find patterns in usage, work out which features to improve or where people are giving up. A capture call requires an event name to specify the event. We recommend using [verb] [noun], like `movie played` or `movie updated` to easily identify what your events mean later on. Capture takes a number of optional arguments, which are defined by the `OptionalCaptureArgs` type.
 
     Examples:
         ```python
         # Context and capture usage
-        from posthog import new_context, identify_context, tag_context, capture
+        from hanzo_insights import new_context, identify_context, tag_context, capture
         # Enter a new context (e.g. a request/response cycle, an instance of a background job, etc)
         with new_context():
             # Associate this context with some user, by distinct_id
@@ -273,7 +312,7 @@ def capture(event: str, **kwargs: Unpack[OptionalCaptureArgs]) -> Optional[str]:
         ```
         ```python
         # Set event properties
-        from posthog import capture
+        from hanzo_insights import capture
         capture(
             "user_signed_up",
             distinct_id="distinct_id_of_the_user",
@@ -300,7 +339,7 @@ def set(**kwargs: Unpack[OptionalSetArgs]) -> Optional[str]:
     Examples:
         ```python
         # Set person properties
-        from posthog import capture
+        from hanzo_insights import capture
         capture(
             'distinct_id',
             event='event_name',
@@ -327,7 +366,7 @@ def set_once(**kwargs: Unpack[OptionalSetArgs]) -> Optional[str]:
     Examples:
         ```python
         # Set property once
-        from posthog import capture
+        from hanzo_insights import capture
         capture(
             'distinct_id',
             event='event_name',
@@ -367,7 +406,7 @@ def group_identify(
     Examples:
         ```python
         # Group identify
-        from posthog import group_identify
+        from hanzo_insights import group_identify
         group_identify('company', 'company_id_in_your_db', {
             'name': 'Awesome Inc.',
             'employees': 11
@@ -412,7 +451,7 @@ def alias(
     Examples:
         ```python
         # Alias user
-        from posthog import alias
+        from hanzo_insights import alias
         alias(previous_id='distinct_id', distinct_id='alias_id')
         ```
     Category:
@@ -440,12 +479,12 @@ def capture_exception(
         exception: The exception to capture. If not provided, the current exception is captured via `sys.exc_info()`
 
     Details:
-        Capture exception is idempotent - if it is called twice with the same exception instance, only a occurrence will be tracked in posthog. This is because, generally, contexts will cause exceptions to be captured automatically. However, to ensure you track an exception, if you catch and do not re-raise it, capturing it manually is recommended, unless you are certain it will have crossed a context boundary (e.g. by existing a `with posthog.new_context():` block already). If the passed exception was raised and caught, the captured stack trace will consist of every frame between where the exception was raised and the point at which it is captured (the "traceback"). If the passed exception was never raised, e.g. if you call `posthog.capture_exception(ValueError("Some Error"))`, the stack trace captured will be the full stack trace at the moment the exception was captured. Note that heavy use of contexts will lead to truncated stack traces, as the exception will be captured by the context entered most recently, which may not be the point you catch the exception for the final time in your code. It's recommended to use contexts sparingly, for this reason. `capture_exception` takes the same set of optional arguments as `capture`.
+        Capture exception is idempotent - if it is called twice with the same exception instance, only a occurrence will be tracked in hanzo_insights. This is because, generally, contexts will cause exceptions to be captured automatically. However, to ensure you track an exception, if you catch and do not re-raise it, capturing it manually is recommended, unless you are certain it will have crossed a context boundary (e.g. by existing a `with hanzo_insights.new_context():` block already). If the passed exception was raised and caught, the captured stack trace will consist of every frame between where the exception was raised and the point at which it is captured (the "traceback"). If the passed exception was never raised, e.g. if you call `hanzo_insights.capture_exception(ValueError("Some Error"))`, the stack trace captured will be the full stack trace at the moment the exception was captured. Note that heavy use of contexts will lead to truncated stack traces, as the exception will be captured by the context entered most recently, which may not be the point you catch the exception for the final time in your code. It's recommended to use contexts sparingly, for this reason. `capture_exception` takes the same set of optional arguments as `capture`.
 
     Examples:
         ```python
         # Capture exception
-        from posthog import capture_exception
+        from hanzo_insights import capture_exception
         try:
             risky_operation()
         except Exception as e:
@@ -467,6 +506,7 @@ def feature_enabled(
     only_evaluate_locally=False,  # type: bool
     send_feature_flag_events=True,  # type: bool
     disable_geoip=None,  # type: Optional[bool]
+    device_id=None,  # type: Optional[str]
 ):
     # type: (...) -> bool
     """
@@ -483,12 +523,12 @@ def feature_enabled(
         disable_geoip: Whether to disable GeoIP lookup
 
     Details:
-        You can call `posthog.load_feature_flags()` before to make sure you're not doing unexpected requests.
+        You can call `hanzo_insights.load_feature_flags()` before to make sure you're not doing unexpected requests.
 
     Examples:
         ```python
         # Boolean feature flag
-        from posthog import feature_enabled, get_feature_flag_payload
+        from hanzo_insights import feature_enabled, get_feature_flag_payload
         is_my_flag_enabled = feature_enabled('flag-key', 'distinct_id_of_your_user')
         if is_my_flag_enabled:
             matched_flag_payload = get_feature_flag_payload('flag-key', 'distinct_id_of_your_user')
@@ -506,6 +546,7 @@ def feature_enabled(
         only_evaluate_locally=only_evaluate_locally,
         send_feature_flag_events=send_feature_flag_events,
         disable_geoip=disable_geoip,
+        device_id=device_id,
     )
 
 
@@ -518,6 +559,7 @@ def get_feature_flag(
     only_evaluate_locally=False,  # type: bool
     send_feature_flag_events=True,  # type: bool
     disable_geoip=None,  # type: Optional[bool]
+    device_id=None,  # type: Optional[str]
 ) -> Optional[FeatureFlag]:
     """
     Get feature flag variant for users. Used with experiments.
@@ -533,12 +575,12 @@ def get_feature_flag(
         disable_geoip: Whether to disable GeoIP lookup
 
     Details:
-        `groups` are a mapping from group type to group key. So, if you have a group type of "organization" and a group key of "5", you would pass groups={"organization": "5"}. `group_properties` take the format: { group_type_name: { group_properties } }. So, for example, if you have the group type "organization" and the group key "5", with the properties name, and employee count, you'll send these as: group_properties={"organization": {"name": "PostHog", "employees": 11}}.
+        `groups` are a mapping from group type to group key. So, if you have a group type of "organization" and a group key of "5", you would pass groups={"organization": "5"}. `group_properties` take the format: { group_type_name: { group_properties } }. So, for example, if you have the group type "organization" and the group key "5", with the properties name, and employee count, you'll send these as: group_properties={"organization": {"name": "Hanzo", "employees": 11}}.
 
     Examples:
         ```python
         # Multivariate feature flag
-        from posthog import get_feature_flag, get_feature_flag_payload
+        from hanzo_insights import get_feature_flag, get_feature_flag_payload
         enabled_variant = get_feature_flag('flag-key', 'distinct_id_of_your_user')
         if enabled_variant == 'variant-key':
             matched_flag_payload = get_feature_flag_payload('flag-key', 'distinct_id_of_your_user')
@@ -556,6 +598,7 @@ def get_feature_flag(
         only_evaluate_locally=only_evaluate_locally,
         send_feature_flag_events=send_feature_flag_events,
         disable_geoip=disable_geoip,
+        device_id=device_id,
     )
 
 
@@ -566,6 +609,7 @@ def get_all_flags(
     group_properties=None,  # type: Optional[dict]
     only_evaluate_locally=False,  # type: bool
     disable_geoip=None,  # type: Optional[bool]
+    device_id=None,  # type: Optional[str]
 ) -> Optional[dict[str, FeatureFlag]]:
     """
     Get all flags for a given user.
@@ -584,7 +628,7 @@ def get_all_flags(
     Examples:
         ```python
         # All flags for user
-        from posthog import get_all_flags
+        from hanzo_insights import get_all_flags
         get_all_flags('distinct_id_of_your_user')
         ```
     Category:
@@ -598,6 +642,7 @@ def get_all_flags(
         group_properties=group_properties or {},
         only_evaluate_locally=only_evaluate_locally,
         disable_geoip=disable_geoip,
+        device_id=device_id,
     )
 
 
@@ -610,6 +655,7 @@ def get_feature_flag_result(
     only_evaluate_locally=False,
     send_feature_flag_events=True,
     disable_geoip=None,  # type: Optional[bool]
+    device_id=None,  # type: Optional[str]
 ):
     # type: (...) -> Optional[FeatureFlagResult]
     """
@@ -624,7 +670,7 @@ def get_feature_flag_result(
 
     Example:
     ```python
-    result = posthog.get_feature_flag_result('beta-feature', 'distinct_id')
+    result = hanzo_insights.get_feature_flag_result('beta-feature', 'distinct_id')
     if result and result.enabled:
         # Use the variant and payload
         print(f"Variant: {result.variant}")
@@ -641,6 +687,7 @@ def get_feature_flag_result(
         only_evaluate_locally=only_evaluate_locally,
         send_feature_flag_events=send_feature_flag_events,
         disable_geoip=disable_geoip,
+        device_id=device_id,
     )
 
 
@@ -654,6 +701,7 @@ def get_feature_flag_payload(
     only_evaluate_locally=False,
     send_feature_flag_events=True,
     disable_geoip=None,  # type: Optional[bool]
+    device_id=None,  # type: Optional[str]
 ) -> Optional[str]:
     return _proxy(
         "get_feature_flag_payload",
@@ -666,6 +714,7 @@ def get_feature_flag_payload(
         only_evaluate_locally=only_evaluate_locally,
         send_feature_flag_events=send_feature_flag_events,
         disable_geoip=disable_geoip,
+        device_id=device_id,
     )
 
 
@@ -696,6 +745,7 @@ def get_all_flags_and_payloads(
     group_properties=None,  # type: Optional[dict]
     only_evaluate_locally=False,
     disable_geoip=None,  # type: Optional[bool]
+    device_id=None,  # type: Optional[str]
 ) -> FlagsAndPayloads:
     return _proxy(
         "get_all_flags_and_payloads",
@@ -705,6 +755,7 @@ def get_all_flags_and_payloads(
         group_properties=group_properties or {},
         only_evaluate_locally=only_evaluate_locally,
         disable_geoip=disable_geoip,
+        device_id=device_id,
     )
 
 
@@ -717,7 +768,7 @@ def feature_flag_definitions():
 
     Examples:
         ```python
-        from posthog import feature_flag_definitions
+        from hanzo_insights import feature_flag_definitions
         definitions = feature_flag_definitions()
         ```
 
@@ -729,11 +780,11 @@ def feature_flag_definitions():
 
 def load_feature_flags():
     """
-    Load feature flag definitions from PostHog.
+    Load feature flag definitions from the server.
 
     Examples:
         ```python
-        from posthog import load_feature_flags
+        from hanzo_insights import load_feature_flags
         load_feature_flags()
         ```
 
@@ -749,7 +800,7 @@ def flush():
 
     Examples:
         ```python
-        from posthog import flush
+        from hanzo_insights import flush
         flush()
         ```
 
@@ -765,7 +816,7 @@ def join():
 
     Examples:
         ```python
-        from posthog import join
+        from hanzo_insights import join
         join()
         ```
 
@@ -781,7 +832,7 @@ def shutdown():
 
     Examples:
         ```python
-        from posthog import shutdown
+        from hanzo_insights import shutdown
         shutdown()
         ```
 
@@ -837,5 +888,7 @@ def _proxy(method, *args, **kwargs):
     return fn(*args, **kwargs)
 
 
-class Posthog(Client):
+class Insights(Client):
+    """Hanzo Insights client for product analytics."""
+
     pass

hanzo_insights/ai/__init__.py

@@ -0,0 +1,3 @@
+from hanzo_insights.ai.prompts import Prompts
+
+__all__ = ["Prompts"]

hanzo_insights/ai/anthropic/anthropic.py

@@ -10,38 +10,38 @@ import time
 import uuid
 from typing import Any, Dict, List, Optional
 
-from posthog.ai.types import StreamingContentBlock, TokenUsage, ToolInProgress
-from posthog.ai.utils import (
+from hanzo_insights.ai.types import StreamingContentBlock, TokenUsage, ToolInProgress
+from hanzo_insights.ai.utils import (
     call_llm_and_track_usage,
     merge_usage_stats,
 )
-from posthog.ai.anthropic.anthropic_converter import (
+from hanzo_insights.ai.anthropic.anthropic_converter import (
     extract_anthropic_usage_from_event,
     handle_anthropic_content_block_start,
     handle_anthropic_text_delta,
     handle_anthropic_tool_delta,
     finalize_anthropic_tool_input,
 )
-from posthog.ai.sanitization import sanitize_anthropic
-from posthog.client import Client as PostHogClient
-from posthog import setup
+from hanzo_insights.ai.sanitization import sanitize_anthropic
+from hanzo_insights.client import Client as InsightsClient
+from hanzo_insights import setup
 
 
 class Anthropic(anthropic.Anthropic):
     """
-    A wrapper around the Anthropic SDK that automatically sends LLM usage events to PostHog.
+    A wrapper around the Anthropic SDK that automatically sends LLM usage events to Insights.
     """
 
-    _ph_client: PostHogClient
+    _ph_client: InsightsClient
 
-    def __init__(self, posthog_client: Optional[PostHogClient] = None, **kwargs):
+    def __init__(self, insights_client: Optional[InsightsClient] = None, **kwargs):
         """
         Args:
-            posthog_client: PostHog client for tracking usage
+            insights_client: Insights client for tracking usage
             **kwargs: Additional arguments passed to the Anthropic client
         """
         super().__init__(**kwargs)
-        self._ph_client = posthog_client or setup()
+        self._ph_client = insights_client or setup()
         self.messages = WrappedMessages(self)
 
 
@@ -50,46 +50,46 @@ class WrappedMessages(Messages):
 
     def create(
         self,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_trace_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: bool = False,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_trace_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: bool = False,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs: Any,
     ):
         """
-        Create a message using Anthropic's API while tracking usage in PostHog.
+        Create a message using Anthropic's API while tracking usage in Insights.
 
         Args:
-            posthog_distinct_id: Optional ID to associate with the usage event
-            posthog_trace_id: Optional trace UUID for linking events
-            posthog_properties: Optional dictionary of extra properties to include in the event
-            posthog_privacy_mode: Whether to redact sensitive information in tracking
-            posthog_groups: Optional group analytics properties
+            insights_distinct_id: Optional ID to associate with the usage event
+            insights_trace_id: Optional trace UUID for linking events
+            insights_properties: Optional dictionary of extra properties to include in the event
+            insights_privacy_mode: Whether to redact sensitive information in tracking
+            insights_groups: Optional group analytics properties
             **kwargs: Arguments passed to Anthropic's messages.create
         """
 
-        if posthog_trace_id is None:
-            posthog_trace_id = str(uuid.uuid4())
+        if insights_trace_id is None:
+            insights_trace_id = str(uuid.uuid4())
 
         if kwargs.get("stream", False):
             return self._create_streaming(
-                posthog_distinct_id,
-                posthog_trace_id,
-                posthog_properties,
-                posthog_privacy_mode,
-                posthog_groups,
+                insights_distinct_id,
+                insights_trace_id,
+                insights_properties,
+                insights_privacy_mode,
+                insights_groups,
                 **kwargs,
             )
 
         return call_llm_and_track_usage(
-            posthog_distinct_id,
+            insights_distinct_id,
             self._client._ph_client,
             "anthropic",
-            posthog_trace_id,
-            posthog_properties,
-            posthog_privacy_mode,
-            posthog_groups,
+            insights_trace_id,
+            insights_properties,
+            insights_privacy_mode,
+            insights_groups,
             self._client.base_url,
             super().create,
             **kwargs,
@@ -97,32 +97,32 @@ class WrappedMessages(Messages):
 
     def stream(
         self,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_trace_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: bool = False,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_trace_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: bool = False,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs: Any,
     ):
-        if posthog_trace_id is None:
-            posthog_trace_id = str(uuid.uuid4())
+        if insights_trace_id is None:
+            insights_trace_id = str(uuid.uuid4())
 
         return self._create_streaming(
-            posthog_distinct_id,
-            posthog_trace_id,
-            posthog_properties,
-            posthog_privacy_mode,
-            posthog_groups,
+            insights_distinct_id,
+            insights_trace_id,
+            insights_properties,
+            insights_privacy_mode,
+            insights_groups,
             **kwargs,
         )
 
     def _create_streaming(
         self,
-        posthog_distinct_id: Optional[str],
-        posthog_trace_id: Optional[str],
-        posthog_properties: Optional[Dict[str, Any]],
-        posthog_privacy_mode: bool,
-        posthog_groups: Optional[Dict[str, Any]],
+        insights_distinct_id: Optional[str],
+        insights_trace_id: Optional[str],
+        insights_properties: Optional[Dict[str, Any]],
+        insights_privacy_mode: bool,
+        insights_groups: Optional[Dict[str, Any]],
         **kwargs: Any,
     ):
         start_time = time.time()
@@ -188,11 +188,11 @@ class WrappedMessages(Messages):
                 latency = end_time - start_time
 
                 self._capture_streaming_event(
-                    posthog_distinct_id,
-                    posthog_trace_id,
-                    posthog_properties,
-                    posthog_privacy_mode,
-                    posthog_groups,
+                    insights_distinct_id,
+                    insights_trace_id,
+                    insights_properties,
+                    insights_privacy_mode,
+                    insights_groups,
                     kwargs,
                     usage_stats,
                     latency,
@@ -204,23 +204,23 @@ class WrappedMessages(Messages):
 
     def _capture_streaming_event(
         self,
-        posthog_distinct_id: Optional[str],
-        posthog_trace_id: Optional[str],
-        posthog_properties: Optional[Dict[str, Any]],
-        posthog_privacy_mode: bool,
-        posthog_groups: Optional[Dict[str, Any]],
+        insights_distinct_id: Optional[str],
+        insights_trace_id: Optional[str],
+        insights_properties: Optional[Dict[str, Any]],
+        insights_privacy_mode: bool,
+        insights_groups: Optional[Dict[str, Any]],
         kwargs: Dict[str, Any],
         usage_stats: TokenUsage,
         latency: float,
         content_blocks: List[StreamingContentBlock],
         accumulated_content: str,
     ):
-        from posthog.ai.types import StreamingEventData
-        from posthog.ai.anthropic.anthropic_converter import (
+        from hanzo_insights.ai.types import StreamingEventData
+        from hanzo_insights.ai.anthropic.anthropic_converter import (
             format_anthropic_streaming_input,
             format_anthropic_streaming_output_complete,
         )
-        from posthog.ai.utils import capture_streaming_event
+        from hanzo_insights.ai.utils import capture_streaming_event
 
         # Prepare standardized event data
         formatted_input = format_anthropic_streaming_input(kwargs)
@@ -237,11 +237,11 @@ class WrappedMessages(Messages):
             ),
             usage_stats=usage_stats,
             latency=latency,
-            distinct_id=posthog_distinct_id,
-            trace_id=posthog_trace_id,
-            properties=posthog_properties,
-            privacy_mode=posthog_privacy_mode,
-            groups=posthog_groups,
+            distinct_id=insights_distinct_id,
+            trace_id=insights_trace_id,
+            properties=insights_properties,
+            privacy_mode=insights_privacy_mode,
+            groups=insights_groups,
         )
 
         # Use the common capture function

hanzo_insights/ai/anthropic/anthropic_async.py

@@ -10,38 +10,38 @@ import time
 import uuid
 from typing import Any, Dict, List, Optional
 
-from posthog import setup
-from posthog.ai.types import StreamingContentBlock, TokenUsage, ToolInProgress
-from posthog.ai.utils import (
+from hanzo_insights import setup
+from hanzo_insights.ai.types import StreamingContentBlock, TokenUsage, ToolInProgress
+from hanzo_insights.ai.utils import (
     call_llm_and_track_usage_async,
     merge_usage_stats,
 )
-from posthog.ai.anthropic.anthropic_converter import (
+from hanzo_insights.ai.anthropic.anthropic_converter import (
     extract_anthropic_usage_from_event,
     handle_anthropic_content_block_start,
     handle_anthropic_text_delta,
     handle_anthropic_tool_delta,
     finalize_anthropic_tool_input,
 )
-from posthog.ai.sanitization import sanitize_anthropic
-from posthog.client import Client as PostHogClient
+from hanzo_insights.ai.sanitization import sanitize_anthropic
+from hanzo_insights.client import Client as InsightsClient
 
 
 class AsyncAnthropic(anthropic.AsyncAnthropic):
     """
-    An async wrapper around the Anthropic SDK that automatically sends LLM usage events to PostHog.
+    An async wrapper around the Anthropic SDK that automatically sends LLM usage events to Insights.
     """
 
-    _ph_client: PostHogClient
+    _ph_client: InsightsClient
 
-    def __init__(self, posthog_client: Optional[PostHogClient] = None, **kwargs):
+    def __init__(self, insights_client: Optional[InsightsClient] = None, **kwargs):
         """
         Args:
-            posthog_client: PostHog client for tracking usage
+            insights_client: Insights client for tracking usage
             **kwargs: Additional arguments passed to the Anthropic client
         """
         super().__init__(**kwargs)
-        self._ph_client = posthog_client or setup()
+        self._ph_client = insights_client or setup()
         self.messages = AsyncWrappedMessages(self)
 
 
@@ -50,46 +50,46 @@ class AsyncWrappedMessages(AsyncMessages):
 
     async def create(
         self,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_trace_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: bool = False,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_trace_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: bool = False,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs: Any,
     ):
         """
-        Create a message using Anthropic's API while tracking usage in PostHog.
+        Create a message using Anthropic's API while tracking usage in Insights.
 
         Args:
-            posthog_distinct_id: Optional ID to associate with the usage event
-            posthog_trace_id: Optional trace UUID for linking events
-            posthog_properties: Optional dictionary of extra properties to include in the event
-            posthog_privacy_mode: Whether to redact sensitive information in tracking
-            posthog_groups: Optional group analytics properties
+            insights_distinct_id: Optional ID to associate with the usage event
+            insights_trace_id: Optional trace UUID for linking events
+            insights_properties: Optional dictionary of extra properties to include in the event
+            insights_privacy_mode: Whether to redact sensitive information in tracking
+            insights_groups: Optional group analytics properties
             **kwargs: Arguments passed to Anthropic's messages.create
         """
 
-        if posthog_trace_id is None:
-            posthog_trace_id = str(uuid.uuid4())
+        if insights_trace_id is None:
+            insights_trace_id = str(uuid.uuid4())
 
         if kwargs.get("stream", False):
             return await self._create_streaming(
-                posthog_distinct_id,
-                posthog_trace_id,
-                posthog_properties,
-                posthog_privacy_mode,
-                posthog_groups,
+                insights_distinct_id,
+                insights_trace_id,
+                insights_properties,
+                insights_privacy_mode,
+                insights_groups,
                 **kwargs,
             )
 
         return await call_llm_and_track_usage_async(
-            posthog_distinct_id,
+            insights_distinct_id,
             self._client._ph_client,
             "anthropic",
-            posthog_trace_id,
-            posthog_properties,
-            posthog_privacy_mode,
-            posthog_groups,
+            insights_trace_id,
+            insights_properties,
+            insights_privacy_mode,
+            insights_groups,
             self._client.base_url,
             super().create,
             **kwargs,
@@ -97,32 +97,32 @@ class AsyncWrappedMessages(AsyncMessages):
 
     async def stream(
         self,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_trace_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: bool = False,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_trace_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: bool = False,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs: Any,
     ):
-        if posthog_trace_id is None:
-            posthog_trace_id = str(uuid.uuid4())
+        if insights_trace_id is None:
+            insights_trace_id = str(uuid.uuid4())
 
         return await self._create_streaming(
-            posthog_distinct_id,
-            posthog_trace_id,
-            posthog_properties,
-            posthog_privacy_mode,
-            posthog_groups,
+            insights_distinct_id,
+            insights_trace_id,
+            insights_properties,
+            insights_privacy_mode,
+            insights_groups,
             **kwargs,
         )
 
     async def _create_streaming(
         self,
-        posthog_distinct_id: Optional[str],
-        posthog_trace_id: Optional[str],
-        posthog_properties: Optional[Dict[str, Any]],
-        posthog_privacy_mode: bool,
-        posthog_groups: Optional[Dict[str, Any]],
+        insights_distinct_id: Optional[str],
+        insights_trace_id: Optional[str],
+        insights_properties: Optional[Dict[str, Any]],
+        insights_privacy_mode: bool,
+        insights_groups: Optional[Dict[str, Any]],
         **kwargs: Any,
     ):
         start_time = time.time()
@@ -188,11 +188,11 @@ class AsyncWrappedMessages(AsyncMessages):
                 latency = end_time - start_time
 
                 await self._capture_streaming_event(
-                    posthog_distinct_id,
-                    posthog_trace_id,
-                    posthog_properties,
-                    posthog_privacy_mode,
-                    posthog_groups,
+                    insights_distinct_id,
+                    insights_trace_id,
+                    insights_properties,
+                    insights_privacy_mode,
+                    insights_groups,
                     kwargs,
                     usage_stats,
                     latency,
@@ -204,23 +204,23 @@ class AsyncWrappedMessages(AsyncMessages):
 
     async def _capture_streaming_event(
         self,
-        posthog_distinct_id: Optional[str],
-        posthog_trace_id: Optional[str],
-        posthog_properties: Optional[Dict[str, Any]],
-        posthog_privacy_mode: bool,
-        posthog_groups: Optional[Dict[str, Any]],
+        insights_distinct_id: Optional[str],
+        insights_trace_id: Optional[str],
+        insights_properties: Optional[Dict[str, Any]],
+        insights_privacy_mode: bool,
+        insights_groups: Optional[Dict[str, Any]],
         kwargs: Dict[str, Any],
         usage_stats: TokenUsage,
         latency: float,
         content_blocks: List[StreamingContentBlock],
         accumulated_content: str,
     ):
-        from posthog.ai.types import StreamingEventData
-        from posthog.ai.anthropic.anthropic_converter import (
+        from hanzo_insights.ai.types import StreamingEventData
+        from hanzo_insights.ai.anthropic.anthropic_converter import (
             format_anthropic_streaming_input,
             format_anthropic_streaming_output_complete,
         )
-        from posthog.ai.utils import capture_streaming_event
+        from hanzo_insights.ai.utils import capture_streaming_event
 
         # Prepare standardized event data
         formatted_input = format_anthropic_streaming_input(kwargs)
@@ -237,11 +237,11 @@ class AsyncWrappedMessages(AsyncMessages):
             ),
             usage_stats=usage_stats,
             latency=latency,
-            distinct_id=posthog_distinct_id,
-            trace_id=posthog_trace_id,
-            properties=posthog_properties,
-            privacy_mode=posthog_privacy_mode,
-            groups=posthog_groups,
+            distinct_id=insights_distinct_id,
+            trace_id=insights_trace_id,
+            properties=insights_properties,
+            privacy_mode=insights_privacy_mode,
+            groups=insights_groups,
         )
 
         # Use the common capture function

hanzo_insights/ai/anthropic/anthropic_converter.py

@@ -2,13 +2,13 @@
 Anthropic-specific conversion utilities.
 
 This module handles the conversion of Anthropic API responses and inputs
-into standardized formats for PostHog tracking.
+into standardized formats for Insights tracking.
 """
 
 import json
 from typing import Any, Dict, List, Optional, Tuple
 
-from posthog.ai.types import (
+from hanzo_insights.ai.types import (
     FormattedContentItem,
     FormattedFunctionCall,
     FormattedMessage,
@@ -17,6 +17,7 @@ from posthog.ai.types import (
     TokenUsage,
     ToolInProgress,
 )
+from hanzo_insights.ai.utils import serialize_raw_usage
 
 
 def format_anthropic_response(response: Any) -> List[FormattedMessage]:
@@ -221,6 +222,12 @@ def extract_anthropic_usage_from_response(response: Any) -> TokenUsage:
     if web_search_count > 0:
         result["web_search_count"] = web_search_count
 
+    # Capture raw usage metadata for backend processing
+    # Serialize to dict here in the converter (not in utils)
+    serialized = serialize_raw_usage(response.usage)
+    if serialized:
+        result["raw_usage"] = serialized
+
     return result
 
 
@@ -247,6 +254,11 @@ def extract_anthropic_usage_from_event(event: Any) -> TokenUsage:
             usage["cache_read_input_tokens"] = getattr(
                 event.message.usage, "cache_read_input_tokens", 0
             )
+            # Capture raw usage metadata for backend processing
+            # Serialize to dict here in the converter (not in utils)
+            serialized = serialize_raw_usage(event.message.usage)
+            if serialized:
+                usage["raw_usage"] = serialized
 
     # Handle usage stats from message_delta event
     if hasattr(event, "usage") and event.usage:
@@ -262,6 +274,12 @@ def extract_anthropic_usage_from_event(event: Any) -> TokenUsage:
                 if web_search_count > 0:
                     usage["web_search_count"] = web_search_count
 
+        # Capture raw usage metadata for backend processing
+        # Serialize to dict here in the converter (not in utils)
+        serialized = serialize_raw_usage(event.usage)
+        if serialized:
+            usage["raw_usage"] = serialized
+
     return usage
 
 
@@ -407,9 +425,9 @@ def format_anthropic_streaming_input(kwargs: Dict[str, Any]) -> Any:
         kwargs: Keyword arguments passed to Anthropic API
 
     Returns:
-        Formatted input ready for PostHog tracking
+        Formatted input ready for Insights tracking
     """
-    from posthog.ai.utils import merge_system_prompt
+    from hanzo_insights.ai.utils import merge_system_prompt
 
     return merge_system_prompt(kwargs, "anthropic")
 
@@ -427,7 +445,7 @@ def format_anthropic_streaming_output_complete(
         accumulated_content: Raw accumulated text content as fallback
 
     Returns:
-        Formatted messages ready for PostHog tracking
+        Formatted messages ready for Insights tracking
     """
     formatted_content = format_anthropic_streaming_content(content_blocks)
 

hanzo_insights/ai/anthropic/anthropic_providers.py

@@ -7,59 +7,59 @@ except ImportError:
 
 from typing import Optional
 
-from posthog.ai.anthropic.anthropic import WrappedMessages
-from posthog.ai.anthropic.anthropic_async import AsyncWrappedMessages
-from posthog.client import Client as PostHogClient
-from posthog import setup
+from hanzo_insights.ai.anthropic.anthropic import WrappedMessages
+from hanzo_insights.ai.anthropic.anthropic_async import AsyncWrappedMessages
+from hanzo_insights.client import Client as InsightsClient
+from hanzo_insights import setup
 
 
 class AnthropicBedrock(anthropic.AnthropicBedrock):
     """
-    A wrapper around the Anthropic Bedrock SDK that automatically sends LLM usage events to PostHog.
+    A wrapper around the Anthropic Bedrock SDK that automatically sends LLM usage events to Insights.
     """
 
-    _ph_client: PostHogClient
+    _ph_client: InsightsClient
 
-    def __init__(self, posthog_client: Optional[PostHogClient] = None, **kwargs):
+    def __init__(self, insights_client: Optional[InsightsClient] = None, **kwargs):
         super().__init__(**kwargs)
-        self._ph_client = posthog_client or setup()
+        self._ph_client = insights_client or setup()
         self.messages = WrappedMessages(self)
 
 
 class AsyncAnthropicBedrock(anthropic.AsyncAnthropicBedrock):
     """
-    A wrapper around the Anthropic Bedrock SDK that automatically sends LLM usage events to PostHog.
+    A wrapper around the Anthropic Bedrock SDK that automatically sends LLM usage events to Insights.
     """
 
-    _ph_client: PostHogClient
+    _ph_client: InsightsClient
 
-    def __init__(self, posthog_client: Optional[PostHogClient] = None, **kwargs):
+    def __init__(self, insights_client: Optional[InsightsClient] = None, **kwargs):
         super().__init__(**kwargs)
-        self._ph_client = posthog_client or setup()
+        self._ph_client = insights_client or setup()
         self.messages = AsyncWrappedMessages(self)
 
 
 class AnthropicVertex(anthropic.AnthropicVertex):
     """
-    A wrapper around the Anthropic Vertex SDK that automatically sends LLM usage events to PostHog.
+    A wrapper around the Anthropic Vertex SDK that automatically sends LLM usage events to Insights.
     """
 
-    _ph_client: PostHogClient
+    _ph_client: InsightsClient
 
-    def __init__(self, posthog_client: Optional[PostHogClient] = None, **kwargs):
+    def __init__(self, insights_client: Optional[InsightsClient] = None, **kwargs):
         super().__init__(**kwargs)
-        self._ph_client = posthog_client or setup()
+        self._ph_client = insights_client or setup()
         self.messages = WrappedMessages(self)
 
 
 class AsyncAnthropicVertex(anthropic.AsyncAnthropicVertex):
     """
-    A wrapper around the Anthropic Vertex SDK that automatically sends LLM usage events to PostHog.
+    A wrapper around the Anthropic Vertex SDK that automatically sends LLM usage events to Insights.
     """
 
-    _ph_client: PostHogClient
+    _ph_client: InsightsClient
 
-    def __init__(self, posthog_client: Optional[PostHogClient] = None, **kwargs):
+    def __init__(self, insights_client: Optional[InsightsClient] = None, **kwargs):
         super().__init__(**kwargs)
-        self._ph_client = posthog_client or setup()
+        self._ph_client = insights_client or setup()
         self.messages = AsyncWrappedMessages(self)

hanzo_insights/ai/gemini/gemini.py

@@ -3,8 +3,8 @@ import time
 import uuid
 from typing import Any, Dict, Optional
 
-from posthog.ai.types import TokenUsage, StreamingEventData
-from posthog.ai.utils import merge_system_prompt
+from hanzo_insights.ai.types import TokenUsage, StreamingEventData
+from hanzo_insights.ai.utils import merge_system_prompt
 
 try:
     from google import genai
@@ -13,40 +13,40 @@ except ImportError:
         "Please install the Google Gemini SDK to use this feature: 'pip install google-genai'"
     )
 
-from posthog import setup
-from posthog.ai.utils import (
+from hanzo_insights import setup
+from hanzo_insights.ai.utils import (
     call_llm_and_track_usage,
     capture_streaming_event,
     merge_usage_stats,
 )
-from posthog.ai.gemini.gemini_converter import (
+from hanzo_insights.ai.gemini.gemini_converter import (
     extract_gemini_usage_from_chunk,
     extract_gemini_content_from_chunk,
     format_gemini_streaming_output,
 )
-from posthog.ai.sanitization import sanitize_gemini
-from posthog.client import Client as PostHogClient
+from hanzo_insights.ai.sanitization import sanitize_gemini
+from hanzo_insights.client import Client as InsightsClient
 
 
 class Client:
     """
-    A drop-in replacement for genai.Client that automatically sends LLM usage events to PostHog.
+    A drop-in replacement for genai.Client that automatically sends LLM usage events to Insights.
 
     Usage:
         client = Client(
             api_key="your_api_key",
-            posthog_client=posthog_client,
-            posthog_distinct_id="default_user",  # Optional defaults
-            posthog_properties={"team": "ai"}    # Optional defaults
+            insights_client=insights_client,
+            insights_distinct_id="default_user",  # Optional defaults
+            insights_properties={"team": "ai"}    # Optional defaults
         )
         response = client.models.generate_content(
             model="gemini-2.0-flash",
             contents=["Hello world"],
-            posthog_distinct_id="specific_user"  # Override default
+            insights_distinct_id="specific_user"  # Override default
         )
     """
 
-    _ph_client: PostHogClient
+    _ph_client: InsightsClient
 
     def __init__(
         self,
@@ -57,11 +57,11 @@ class Client:
         location: Optional[str] = None,
         debug_config: Optional[Any] = None,
         http_options: Optional[Any] = None,
-        posthog_client: Optional[PostHogClient] = None,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: bool = False,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_client: Optional[InsightsClient] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: bool = False,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs,
     ):
         """
@@ -73,18 +73,18 @@ class Client:
             location: GCP location for Vertex AI
             debug_config: Debug configuration for the client
             http_options: HTTP options for the client
-            posthog_client: PostHog client for tracking usage
-            posthog_distinct_id: Default distinct ID for all calls (can be overridden per call)
-            posthog_properties: Default properties for all calls (can be overridden per call)
-            posthog_privacy_mode: Default privacy mode for all calls (can be overridden per call)
-            posthog_groups: Default groups for all calls (can be overridden per call)
+            insights_client: Insights client for tracking usage
+            insights_distinct_id: Default distinct ID for all calls (can be overridden per call)
+            insights_properties: Default properties for all calls (can be overridden per call)
+            insights_privacy_mode: Default privacy mode for all calls (can be overridden per call)
+            insights_groups: Default groups for all calls (can be overridden per call)
             **kwargs: Additional arguments (for future compatibility)
         """
 
-        self._ph_client = posthog_client or setup()
+        self._ph_client = insights_client or setup()
 
         if self._ph_client is None:
-            raise ValueError("posthog_client is required for PostHog tracking")
+            raise ValueError("insights_client is required for Insights tracking")
 
         self.models = Models(
             api_key=api_key,
@@ -94,21 +94,21 @@ class Client:
             location=location,
             debug_config=debug_config,
             http_options=http_options,
-            posthog_client=self._ph_client,
-            posthog_distinct_id=posthog_distinct_id,
-            posthog_properties=posthog_properties,
-            posthog_privacy_mode=posthog_privacy_mode,
-            posthog_groups=posthog_groups,
+            insights_client=self._ph_client,
+            insights_distinct_id=insights_distinct_id,
+            insights_properties=insights_properties,
+            insights_privacy_mode=insights_privacy_mode,
+            insights_groups=insights_groups,
             **kwargs,
         )
 
 
 class Models:
     """
-    Models interface that mimics genai.Client().models with PostHog tracking.
+    Models interface that mimics genai.Client().models with Insights tracking.
     """
 
-    _ph_client: PostHogClient  # Not None after __init__ validation
+    _ph_client: InsightsClient  # Not None after __init__ validation
 
     def __init__(
         self,
@@ -119,11 +119,11 @@ class Models:
         location: Optional[str] = None,
         debug_config: Optional[Any] = None,
         http_options: Optional[Any] = None,
-        posthog_client: Optional[PostHogClient] = None,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: bool = False,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_client: Optional[InsightsClient] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: bool = False,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs,
     ):
         """
@@ -135,24 +135,24 @@ class Models:
             location: GCP location for Vertex AI
             debug_config: Debug configuration for the client
             http_options: HTTP options for the client
-            posthog_client: PostHog client for tracking usage
-            posthog_distinct_id: Default distinct ID for all calls
-            posthog_properties: Default properties for all calls
-            posthog_privacy_mode: Default privacy mode for all calls
-            posthog_groups: Default groups for all calls
+            insights_client: Insights client for tracking usage
+            insights_distinct_id: Default distinct ID for all calls
+            insights_properties: Default properties for all calls
+            insights_privacy_mode: Default privacy mode for all calls
+            insights_groups: Default groups for all calls
             **kwargs: Additional arguments (for future compatibility)
         """
 
-        self._ph_client = posthog_client or setup()
+        self._ph_client = insights_client or setup()
 
         if self._ph_client is None:
-            raise ValueError("posthog_client is required for PostHog tracking")
+            raise ValueError("insights_client is required for Insights tracking")
 
-        # Store default PostHog settings
-        self._default_distinct_id = posthog_distinct_id
-        self._default_properties = posthog_properties or {}
-        self._default_privacy_mode = posthog_privacy_mode
-        self._default_groups = posthog_groups
+        # Store default Insights settings
+        self._default_distinct_id = insights_distinct_id
+        self._default_properties = insights_properties or {}
+        self._default_privacy_mode = insights_privacy_mode
+        self._default_groups = insights_groups
 
         # Build genai.Client arguments
         client_args: Dict[str, Any] = {}
@@ -196,7 +196,7 @@ class Models:
         self._client = genai.Client(**client_args)
         self._base_url = "https://generativelanguage.googleapis.com"
 
-    def _merge_posthog_params(
+    def _merge_insights_params(
         self,
         call_distinct_id: Optional[str],
         call_trace_id: Optional[str],
@@ -204,7 +204,7 @@ class Models:
         call_privacy_mode: Optional[bool],
         call_groups: Optional[Dict[str, Any]],
     ):
-        """Merge call-level PostHog parameters with client defaults."""
+        """Merge call-level Insights parameters with client defaults."""
 
         # Use call-level values if provided, otherwise fall back to defaults
         distinct_id = (
@@ -234,38 +234,38 @@ class Models:
         self,
         model: str,
         contents,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_trace_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: Optional[bool] = None,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_trace_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: Optional[bool] = None,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs: Any,
     ):
         """
-        Generate content using Gemini's API while tracking usage in PostHog.
+        Generate content using Gemini's API while tracking usage in Insights.
 
         This method signature exactly matches genai.Client().models.generate_content()
-        with additional PostHog tracking parameters.
+        with additional Insights tracking parameters.
 
         Args:
             model: The model to use (e.g., 'gemini-2.0-flash')
             contents: The input content for generation
-            posthog_distinct_id: ID to associate with the usage event (overrides client default)
-            posthog_trace_id: Trace UUID for linking events (auto-generated if not provided)
-            posthog_properties: Extra properties to include in the event (merged with client defaults)
-            posthog_privacy_mode: Whether to redact sensitive information (overrides client default)
-            posthog_groups: Group analytics properties (overrides client default)
+            insights_distinct_id: ID to associate with the usage event (overrides client default)
+            insights_trace_id: Trace UUID for linking events (auto-generated if not provided)
+            insights_properties: Extra properties to include in the event (merged with client defaults)
+            insights_privacy_mode: Whether to redact sensitive information (overrides client default)
+            insights_groups: Group analytics properties (overrides client default)
             **kwargs: Arguments passed to Gemini's generate_content
         """
 
-        # Merge PostHog parameters
+        # Merge Insights parameters
         distinct_id, trace_id, properties, privacy_mode, groups = (
-            self._merge_posthog_params(
-                posthog_distinct_id,
-                posthog_trace_id,
-                posthog_properties,
-                posthog_privacy_mode,
-                posthog_groups,
+            self._merge_insights_params(
+                insights_distinct_id,
+                insights_trace_id,
+                insights_properties,
+                insights_privacy_mode,
+                insights_groups,
             )
         )
 
@@ -380,7 +380,7 @@ class Models:
         capture_streaming_event(self._ph_client, event_data)
 
     def _format_input(self, contents, **kwargs):
-        """Format input contents for PostHog tracking"""
+        """Format input contents for Insights tracking"""
 
         # Create kwargs dict with contents for merge_system_prompt
         input_kwargs = {"contents": contents, **kwargs}
@@ -390,21 +390,21 @@ class Models:
         self,
         model: str,
         contents,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_trace_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: Optional[bool] = None,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_trace_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: Optional[bool] = None,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs: Any,
     ):
-        # Merge PostHog parameters
+        # Merge Insights parameters
         distinct_id, trace_id, properties, privacy_mode, groups = (
-            self._merge_posthog_params(
-                posthog_distinct_id,
-                posthog_trace_id,
-                posthog_properties,
-                posthog_privacy_mode,
-                posthog_groups,
+            self._merge_insights_params(
+                insights_distinct_id,
+                insights_trace_id,
+                insights_properties,
+                insights_privacy_mode,
+                insights_groups,
             )
         )
 

hanzo_insights/ai/gemini/gemini_async.py

@@ -3,8 +3,8 @@ import time
 import uuid
 from typing import Any, Dict, Optional
 
-from posthog.ai.types import TokenUsage, StreamingEventData
-from posthog.ai.utils import merge_system_prompt
+from hanzo_insights.ai.types import TokenUsage, StreamingEventData
+from hanzo_insights.ai.utils import merge_system_prompt
 
 try:
     from google import genai
@@ -13,40 +13,40 @@ except ImportError:
         "Please install the Google Gemini SDK to use this feature: 'pip install google-genai'"
     )
 
-from posthog import setup
-from posthog.ai.utils import (
+from hanzo_insights import setup
+from hanzo_insights.ai.utils import (
     call_llm_and_track_usage_async,
     capture_streaming_event,
     merge_usage_stats,
 )
-from posthog.ai.gemini.gemini_converter import (
+from hanzo_insights.ai.gemini.gemini_converter import (
     extract_gemini_usage_from_chunk,
     extract_gemini_content_from_chunk,
     format_gemini_streaming_output,
 )
-from posthog.ai.sanitization import sanitize_gemini
-from posthog.client import Client as PostHogClient
+from hanzo_insights.ai.sanitization import sanitize_gemini
+from hanzo_insights.client import Client as InsightsClient
 
 
 class AsyncClient:
     """
-    An async drop-in replacement for genai.Client that automatically sends LLM usage events to PostHog.
+    An async drop-in replacement for genai.Client that automatically sends LLM usage events to Insights.
 
     Usage:
         client = AsyncClient(
             api_key="your_api_key",
-            posthog_client=posthog_client,
-            posthog_distinct_id="default_user",  # Optional defaults
-            posthog_properties={"team": "ai"}    # Optional defaults
+            insights_client=insights_client,
+            insights_distinct_id="default_user",  # Optional defaults
+            insights_properties={"team": "ai"}    # Optional defaults
         )
         response = await client.models.generate_content(
             model="gemini-2.0-flash",
             contents=["Hello world"],
-            posthog_distinct_id="specific_user"  # Override default
+            insights_distinct_id="specific_user"  # Override default
         )
     """
 
-    _ph_client: PostHogClient
+    _ph_client: InsightsClient
 
     def __init__(
         self,
@@ -57,11 +57,11 @@ class AsyncClient:
         location: Optional[str] = None,
         debug_config: Optional[Any] = None,
         http_options: Optional[Any] = None,
-        posthog_client: Optional[PostHogClient] = None,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: bool = False,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_client: Optional[InsightsClient] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: bool = False,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs,
     ):
         """
@@ -73,18 +73,18 @@ class AsyncClient:
             location: GCP location for Vertex AI
             debug_config: Debug configuration for the client
             http_options: HTTP options for the client
-            posthog_client: PostHog client for tracking usage
-            posthog_distinct_id: Default distinct ID for all calls (can be overridden per call)
-            posthog_properties: Default properties for all calls (can be overridden per call)
-            posthog_privacy_mode: Default privacy mode for all calls (can be overridden per call)
-            posthog_groups: Default groups for all calls (can be overridden per call)
+            insights_client: Insights client for tracking usage
+            insights_distinct_id: Default distinct ID for all calls (can be overridden per call)
+            insights_properties: Default properties for all calls (can be overridden per call)
+            insights_privacy_mode: Default privacy mode for all calls (can be overridden per call)
+            insights_groups: Default groups for all calls (can be overridden per call)
             **kwargs: Additional arguments (for future compatibility)
         """
 
-        self._ph_client = posthog_client or setup()
+        self._ph_client = insights_client or setup()
 
         if self._ph_client is None:
-            raise ValueError("posthog_client is required for PostHog tracking")
+            raise ValueError("insights_client is required for Insights tracking")
 
         self.models = AsyncModels(
             api_key=api_key,
@@ -94,21 +94,21 @@ class AsyncClient:
             location=location,
             debug_config=debug_config,
             http_options=http_options,
-            posthog_client=self._ph_client,
-            posthog_distinct_id=posthog_distinct_id,
-            posthog_properties=posthog_properties,
-            posthog_privacy_mode=posthog_privacy_mode,
-            posthog_groups=posthog_groups,
+            insights_client=self._ph_client,
+            insights_distinct_id=insights_distinct_id,
+            insights_properties=insights_properties,
+            insights_privacy_mode=insights_privacy_mode,
+            insights_groups=insights_groups,
             **kwargs,
         )
 
 
 class AsyncModels:
     """
-    Async Models interface that mimics genai.Client().aio.models with PostHog tracking.
+    Async Models interface that mimics genai.Client().aio.models with Insights tracking.
     """
 
-    _ph_client: PostHogClient  # Not None after __init__ validation
+    _ph_client: InsightsClient  # Not None after __init__ validation
 
     def __init__(
         self,
@@ -119,11 +119,11 @@ class AsyncModels:
         location: Optional[str] = None,
         debug_config: Optional[Any] = None,
         http_options: Optional[Any] = None,
-        posthog_client: Optional[PostHogClient] = None,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: bool = False,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_client: Optional[InsightsClient] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: bool = False,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs,
     ):
         """
@@ -135,24 +135,24 @@ class AsyncModels:
             location: GCP location for Vertex AI
             debug_config: Debug configuration for the client
             http_options: HTTP options for the client
-            posthog_client: PostHog client for tracking usage
-            posthog_distinct_id: Default distinct ID for all calls
-            posthog_properties: Default properties for all calls
-            posthog_privacy_mode: Default privacy mode for all calls
-            posthog_groups: Default groups for all calls
+            insights_client: Insights client for tracking usage
+            insights_distinct_id: Default distinct ID for all calls
+            insights_properties: Default properties for all calls
+            insights_privacy_mode: Default privacy mode for all calls
+            insights_groups: Default groups for all calls
             **kwargs: Additional arguments (for future compatibility)
         """
 
-        self._ph_client = posthog_client or setup()
+        self._ph_client = insights_client or setup()
 
         if self._ph_client is None:
-            raise ValueError("posthog_client is required for PostHog tracking")
+            raise ValueError("insights_client is required for Insights tracking")
 
-        # Store default PostHog settings
-        self._default_distinct_id = posthog_distinct_id
-        self._default_properties = posthog_properties or {}
-        self._default_privacy_mode = posthog_privacy_mode
-        self._default_groups = posthog_groups
+        # Store default Insights settings
+        self._default_distinct_id = insights_distinct_id
+        self._default_properties = insights_properties or {}
+        self._default_privacy_mode = insights_privacy_mode
+        self._default_groups = insights_groups
 
         # Build genai.Client arguments
         client_args: Dict[str, Any] = {}
@@ -196,7 +196,7 @@ class AsyncModels:
         self._client = genai.Client(**client_args)
         self._base_url = "https://generativelanguage.googleapis.com"
 
-    def _merge_posthog_params(
+    def _merge_insights_params(
         self,
         call_distinct_id: Optional[str],
         call_trace_id: Optional[str],
@@ -204,7 +204,7 @@ class AsyncModels:
         call_privacy_mode: Optional[bool],
         call_groups: Optional[Dict[str, Any]],
     ):
-        """Merge call-level PostHog parameters with client defaults."""
+        """Merge call-level Insights parameters with client defaults."""
 
         # Use call-level values if provided, otherwise fall back to defaults
         distinct_id = (
@@ -234,38 +234,38 @@ class AsyncModels:
         self,
         model: str,
         contents,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_trace_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: Optional[bool] = None,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_trace_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: Optional[bool] = None,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs: Any,
     ):
         """
-        Generate content using Gemini's API while tracking usage in PostHog.
+        Generate content using Gemini's API while tracking usage in Insights.
 
         This method signature exactly matches genai.Client().aio.models.generate_content()
-        with additional PostHog tracking parameters.
+        with additional Insights tracking parameters.
 
         Args:
             model: The model to use (e.g., 'gemini-2.0-flash')
             contents: The input content for generation
-            posthog_distinct_id: ID to associate with the usage event (overrides client default)
-            posthog_trace_id: Trace UUID for linking events (auto-generated if not provided)
-            posthog_properties: Extra properties to include in the event (merged with client defaults)
-            posthog_privacy_mode: Whether to redact sensitive information (overrides client default)
-            posthog_groups: Group analytics properties (overrides client default)
+            insights_distinct_id: ID to associate with the usage event (overrides client default)
+            insights_trace_id: Trace UUID for linking events (auto-generated if not provided)
+            insights_properties: Extra properties to include in the event (merged with client defaults)
+            insights_privacy_mode: Whether to redact sensitive information (overrides client default)
+            insights_groups: Group analytics properties (overrides client default)
             **kwargs: Arguments passed to Gemini's generate_content
         """
 
-        # Merge PostHog parameters
+        # Merge Insights parameters
         distinct_id, trace_id, properties, privacy_mode, groups = (
-            self._merge_posthog_params(
-                posthog_distinct_id,
-                posthog_trace_id,
-                posthog_properties,
-                posthog_privacy_mode,
-                posthog_groups,
+            self._merge_insights_params(
+                insights_distinct_id,
+                insights_trace_id,
+                insights_properties,
+                insights_privacy_mode,
+                insights_groups,
             )
         )
 
@@ -383,7 +383,7 @@ class AsyncModels:
         capture_streaming_event(self._ph_client, event_data)
 
     def _format_input(self, contents, **kwargs):
-        """Format input contents for PostHog tracking"""
+        """Format input contents for Insights tracking"""
 
         # Create kwargs dict with contents for merge_system_prompt
         input_kwargs = {"contents": contents, **kwargs}
@@ -393,21 +393,21 @@ class AsyncModels:
         self,
         model: str,
         contents,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_trace_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: Optional[bool] = None,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_trace_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: Optional[bool] = None,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs: Any,
     ):
-        # Merge PostHog parameters
+        # Merge Insights parameters
         distinct_id, trace_id, properties, privacy_mode, groups = (
-            self._merge_posthog_params(
-                posthog_distinct_id,
-                posthog_trace_id,
-                posthog_properties,
-                posthog_privacy_mode,
-                posthog_groups,
+            self._merge_insights_params(
+                insights_distinct_id,
+                insights_trace_id,
+                insights_properties,
+                insights_privacy_mode,
+                insights_groups,
             )
         )
 

hanzo_insights/ai/gemini/gemini_converter.py

@@ -2,16 +2,17 @@
 Gemini-specific conversion utilities.
 
 This module handles the conversion of Gemini API responses and inputs
-into standardized formats for PostHog tracking.
+into standardized formats for Insights tracking.
 """
 
 from typing import Any, Dict, List, Optional, TypedDict, Union
 
-from posthog.ai.types import (
+from hanzo_insights.ai.types import (
     FormattedContentItem,
     FormattedMessage,
     TokenUsage,
 )
+from hanzo_insights.ai.utils import serialize_raw_usage
 
 
 class GeminiPart(TypedDict, total=False):
@@ -348,7 +349,7 @@ def format_gemini_input_with_system(
     if system_instruction is not None:
         has_system = any(msg.get("role") == "system" for msg in formatted_messages)
         if not has_system:
-            from posthog.ai.types import FormattedMessage
+            from hanzo_insights.ai.types import FormattedMessage
 
             system_message: FormattedMessage = {
                 "role": "system",
@@ -361,7 +362,7 @@ def format_gemini_input_with_system(
 
 def format_gemini_input(contents: Any) -> List[FormattedMessage]:
     """
-    Format Gemini input contents into standardized message format for PostHog tracking.
+    Format Gemini input contents into standardized message format for Insights tracking.
 
     This function handles various input formats:
     - String inputs
@@ -487,6 +488,12 @@ def _extract_usage_from_metadata(metadata: Any) -> TokenUsage:
         if reasoning_tokens and reasoning_tokens > 0:
             usage["reasoning_tokens"] = reasoning_tokens
 
+    # Capture raw usage metadata for backend processing
+    # Serialize to dict here in the converter (not in utils)
+    serialized = serialize_raw_usage(metadata)
+    if serialized:
+        usage["raw_usage"] = serialized
+
     return usage
 
 

hanzo_insights/ai/langchain/callbacks.py

@@ -22,8 +22,8 @@ from uuid import UUID
 
 try:
     # LangChain 1.0+ and modern 0.x with langchain-core
-    from langchain_core.callbacks.base import BaseCallbackHandler
     from langchain_core.agents import AgentAction, AgentFinish
+    from langchain_core.callbacks.base import BaseCallbackHandler
 except (ImportError, ModuleNotFoundError):
     # Fallback for older LangChain versions
     from langchain.callbacks.base import BaseCallbackHandler
@@ -35,18 +35,18 @@ from langchain_core.messages import (
     FunctionMessage,
     HumanMessage,
     SystemMessage,
-    ToolMessage,
     ToolCall,
+    ToolMessage,
 )
 from langchain_core.outputs import ChatGeneration, LLMResult
 from pydantic import BaseModel
 
-from posthog import setup
-from posthog.ai.utils import get_model_params, with_privacy_mode
-from posthog.ai.sanitization import sanitize_langchain
-from posthog.client import Client
+from hanzo_insights import setup
+from hanzo_insights.ai.sanitization import sanitize_langchain
+from hanzo_insights.ai.utils import get_model_params, with_privacy_mode
+from hanzo_insights.client import Client
 
-log = logging.getLogger("posthog")
+log = logging.getLogger("hanzo_insights")
 
 
 @dataclass
@@ -79,8 +79,8 @@ class GenerationMetadata(SpanMetadata):
     """Base URL of the provider's API used in the run."""
     tools: Optional[List[Dict[str, Any]]] = None
     """Tools provided to the model."""
-    posthog_properties: Optional[Dict[str, Any]] = None
-    """PostHog properties of the run."""
+    insights_properties: Optional[Dict[str, Any]] = None
+    """Insights properties of the run."""
 
 
 RunMetadata = Union[SpanMetadata, GenerationMetadata]
@@ -89,11 +89,11 @@ RunMetadataStorage = Dict[UUID, RunMetadata]
 
 class CallbackHandler(BaseCallbackHandler):
     """
-    The PostHog LLM observability callback handler for LangChain.
+    The Insights LLM observability callback handler for LangChain.
     """
 
     _ph_client: Client
-    """PostHog client instance."""
+    """Insights client instance."""
 
     _distinct_id: Optional[Union[str, int, UUID]]
     """Distinct ID of the user to associate the trace with."""
@@ -131,12 +131,12 @@ class CallbackHandler(BaseCallbackHandler):
     ):
         """
         Args:
-            client: PostHog client instance.
+            client: Insights client instance.
             distinct_id: Optional distinct ID of the user to associate the trace with.
             trace_id: Optional trace ID to use for the event.
             properties: Optional additional metadata to use for the trace.
             privacy_mode: Whether to redact the input and output of the trace.
-            groups: Optional additional PostHog groups to use for the trace.
+            groups: Optional additional Insights groups to use for the trace.
         """
         self._ph_client = client or setup()
         self._distinct_id = distinct_id
@@ -423,7 +423,7 @@ class CallbackHandler(BaseCallbackHandler):
             if provider := metadata.get("ls_provider"):
                 generation.provider = provider
 
-            generation.posthog_properties = metadata.get("posthog_properties")
+            generation.insights_properties = metadata.get("insights_properties")
         try:
             base_url = serialized["kwargs"]["openai_api_base"]
             if base_url is not None:
@@ -506,6 +506,14 @@ class CallbackHandler(BaseCallbackHandler):
         if isinstance(outputs, BaseException):
             event_properties["$ai_error"] = _stringify_exception(outputs)
             event_properties["$ai_is_error"] = True
+            event_properties = _capture_exception_and_update_properties(
+                self._ph_client,
+                outputs,
+                self._distinct_id,
+                self._groups,
+                event_properties,
+            )
+
         elif outputs is not None:
             event_properties["$ai_output_state"] = with_privacy_mode(
                 self._ph_client, self._privacy_mode, outputs
@@ -570,16 +578,30 @@ class CallbackHandler(BaseCallbackHandler):
             "$ai_framework": "langchain",
         }
 
-        if isinstance(run.posthog_properties, dict):
-            event_properties.update(run.posthog_properties)
+        if isinstance(run.insights_properties, dict):
+            event_properties.update(run.insights_properties)
 
         if run.tools:
             event_properties["$ai_tools"] = run.tools
 
+        if self._properties:
+            event_properties.update(self._properties)
+
+        if self._distinct_id is None:
+            event_properties["$process_person_profile"] = False
+
         if isinstance(output, BaseException):
             event_properties["$ai_http_status"] = _get_http_status(output)
             event_properties["$ai_error"] = _stringify_exception(output)
             event_properties["$ai_is_error"] = True
+
+            event_properties = _capture_exception_and_update_properties(
+                self._ph_client,
+                output,
+                self._distinct_id,
+                self._groups,
+                event_properties,
+            )
         else:
             # Add usage
             usage = _parse_usage(output, run.provider, run.model)
@@ -607,12 +629,6 @@ class CallbackHandler(BaseCallbackHandler):
                 self._ph_client, self._privacy_mode, completions
             )
 
-        if self._properties:
-            event_properties.update(self._properties)
-
-        if self._distinct_id is None:
-            event_properties["$process_person_profile"] = False
-
         self._ph_client.capture(
             distinct_id=self._distinct_id or trace_id,
             event="$ai_generation",
@@ -863,6 +879,27 @@ def _parse_usage(
     return llm_usage
 
 
+def _capture_exception_and_update_properties(
+    client: Client,
+    exception: BaseException,
+    distinct_id: Optional[Union[str, int, UUID]],
+    groups: Optional[Dict[str, Any]],
+    event_properties: Dict[str, Any],
+):
+    if client.enable_exception_autocapture:
+        exception_id = client.capture_exception(
+            exception,
+            distinct_id=distinct_id,
+            groups=groups,
+            properties=event_properties,
+        )
+
+        if exception_id:
+            event_properties["$exception_event_id"] = exception_id
+
+    return event_properties
+
+
 def _get_http_status(error: BaseException) -> int:
     # OpenAI: https://github.com/openai/openai-python/blob/main/src/openai/_exceptions.py
     # Anthropic: https://github.com/anthropics/anthropic-sdk-python/blob/main/src/anthropic/_exceptions.py

hanzo_insights/ai/openai/openai.py

@@ -2,7 +2,7 @@ import time
 import uuid
 from typing import Any, Dict, List, Optional
 
-from posthog.ai.types import TokenUsage
+from hanzo_insights.ai.types import TokenUsage
 
 try:
     import openai
@@ -11,40 +11,40 @@ except ImportError:
         "Please install the OpenAI SDK to use this feature: 'pip install openai'"
     )
 
-from posthog.ai.utils import (
+from hanzo_insights.ai.utils import (
     call_llm_and_track_usage,
     extract_available_tool_calls,
     merge_usage_stats,
     with_privacy_mode,
 )
-from posthog.ai.openai.openai_converter import (
+from hanzo_insights.ai.openai.openai_converter import (
     extract_openai_usage_from_chunk,
     extract_openai_content_from_chunk,
     extract_openai_tool_calls_from_chunk,
     accumulate_openai_tool_calls,
 )
-from posthog.ai.sanitization import sanitize_openai, sanitize_openai_response
-from posthog.client import Client as PostHogClient
-from posthog import setup
+from hanzo_insights.ai.sanitization import sanitize_openai, sanitize_openai_response
+from hanzo_insights.client import Client as InsightsClient
+from hanzo_insights import setup
 
 
 class OpenAI(openai.OpenAI):
     """
-    A wrapper around the OpenAI SDK that automatically sends LLM usage events to PostHog.
+    A wrapper around the OpenAI SDK that automatically sends LLM usage events to Insights.
     """
 
-    _ph_client: PostHogClient
+    _ph_client: InsightsClient
 
-    def __init__(self, posthog_client: Optional[PostHogClient] = None, **kwargs):
+    def __init__(self, insights_client: Optional[InsightsClient] = None, **kwargs):
         """
         Args:
             api_key: OpenAI API key.
-            posthog_client: If provided, events will be captured via this client instead of the global `posthog`.
+            insights_client: If provided, events will be captured via this client instead of the global client.
             **openai_config: Any additional keyword args to set on openai (e.g. organization="xxx").
         """
 
         super().__init__(**kwargs)
-        self._ph_client = posthog_client or setup()
+        self._ph_client = insights_client or setup()
 
         # Store original objects after parent initialization (only if they exist)
         self._original_chat = getattr(self, "chat", None)
@@ -67,7 +67,7 @@ class OpenAI(openai.OpenAI):
 
 
 class WrappedResponses:
-    """Wrapper for OpenAI responses that tracks usage in PostHog."""
+    """Wrapper for OpenAI responses that tracks usage in Insights."""
 
     def __init__(self, client: OpenAI, original_responses):
         self._client = client
@@ -79,34 +79,34 @@ class WrappedResponses:
 
     def create(
         self,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_trace_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: bool = False,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_trace_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: bool = False,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs: Any,
     ):
-        if posthog_trace_id is None:
-            posthog_trace_id = str(uuid.uuid4())
+        if insights_trace_id is None:
+            insights_trace_id = str(uuid.uuid4())
 
         if kwargs.get("stream", False):
             return self._create_streaming(
-                posthog_distinct_id,
-                posthog_trace_id,
-                posthog_properties,
-                posthog_privacy_mode,
-                posthog_groups,
+                insights_distinct_id,
+                insights_trace_id,
+                insights_properties,
+                insights_privacy_mode,
+                insights_groups,
                 **kwargs,
             )
 
         return call_llm_and_track_usage(
-            posthog_distinct_id,
+            insights_distinct_id,
             self._client._ph_client,
             "openai",
-            posthog_trace_id,
-            posthog_properties,
-            posthog_privacy_mode,
-            posthog_groups,
+            insights_trace_id,
+            insights_properties,
+            insights_privacy_mode,
+            insights_groups,
             self._client.base_url,
             self._original.create,
             **kwargs,
@@ -114,11 +114,11 @@ class WrappedResponses:
 
     def _create_streaming(
         self,
-        posthog_distinct_id: Optional[str],
-        posthog_trace_id: Optional[str],
-        posthog_properties: Optional[Dict[str, Any]],
-        posthog_privacy_mode: bool,
-        posthog_groups: Optional[Dict[str, Any]],
+        insights_distinct_id: Optional[str],
+        insights_trace_id: Optional[str],
+        insights_properties: Optional[Dict[str, Any]],
+        insights_privacy_mode: bool,
+        insights_groups: Optional[Dict[str, Any]],
         **kwargs: Any,
     ):
         start_time = time.time()
@@ -160,11 +160,11 @@ class WrappedResponses:
                 latency = end_time - start_time
                 output = final_content
                 self._capture_streaming_event(
-                    posthog_distinct_id,
-                    posthog_trace_id,
-                    posthog_properties,
-                    posthog_privacy_mode,
-                    posthog_groups,
+                    insights_distinct_id,
+                    insights_trace_id,
+                    insights_properties,
+                    insights_privacy_mode,
+                    insights_groups,
                     kwargs,
                     usage_stats,
                     latency,
@@ -177,11 +177,11 @@ class WrappedResponses:
 
     def _capture_streaming_event(
         self,
-        posthog_distinct_id: Optional[str],
-        posthog_trace_id: Optional[str],
-        posthog_properties: Optional[Dict[str, Any]],
-        posthog_privacy_mode: bool,
-        posthog_groups: Optional[Dict[str, Any]],
+        insights_distinct_id: Optional[str],
+        insights_trace_id: Optional[str],
+        insights_properties: Optional[Dict[str, Any]],
+        insights_privacy_mode: bool,
+        insights_groups: Optional[Dict[str, Any]],
         kwargs: Dict[str, Any],
         usage_stats: TokenUsage,
         latency: float,
@@ -189,12 +189,12 @@ class WrappedResponses:
         available_tool_calls: Optional[List[Dict[str, Any]]] = None,
         model_from_response: Optional[str] = None,
     ):
-        from posthog.ai.types import StreamingEventData
-        from posthog.ai.openai.openai_converter import (
+        from hanzo_insights.ai.types import StreamingEventData
+        from hanzo_insights.ai.openai.openai_converter import (
             format_openai_streaming_input,
             format_openai_streaming_output,
         )
-        from posthog.ai.utils import capture_streaming_event
+        from hanzo_insights.ai.utils import capture_streaming_event
 
         # Prepare standardized event data
         formatted_input = format_openai_streaming_input(kwargs, "responses")
@@ -212,11 +212,11 @@ class WrappedResponses:
             formatted_output=format_openai_streaming_output(output, "responses"),
             usage_stats=usage_stats,
             latency=latency,
-            distinct_id=posthog_distinct_id,
-            trace_id=posthog_trace_id,
-            properties=posthog_properties,
-            privacy_mode=posthog_privacy_mode,
-            groups=posthog_groups,
+            distinct_id=insights_distinct_id,
+            trace_id=insights_trace_id,
+            properties=insights_properties,
+            privacy_mode=insights_privacy_mode,
+            groups=insights_groups,
         )
 
         # Use the common capture function
@@ -224,35 +224,35 @@ class WrappedResponses:
 
     def parse(
         self,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_trace_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: bool = False,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_trace_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: bool = False,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs: Any,
     ):
         """
-        Parse structured output using OpenAI's 'responses.parse' method, but also track usage in PostHog.
+        Parse structured output using OpenAI's 'responses.parse' method, but also track usage in Insights.
 
         Args:
-            posthog_distinct_id: Optional ID to associate with the usage event.
-            posthog_trace_id: Optional trace UUID for linking events.
-            posthog_properties: Optional dictionary of extra properties to include in the event.
-            posthog_privacy_mode: Whether to anonymize the input and output.
-            posthog_groups: Optional dictionary of groups to associate with the event.
+            insights_distinct_id: Optional ID to associate with the usage event.
+            insights_trace_id: Optional trace UUID for linking events.
+            insights_properties: Optional dictionary of extra properties to include in the event.
+            insights_privacy_mode: Whether to anonymize the input and output.
+            insights_groups: Optional dictionary of groups to associate with the event.
             **kwargs: Any additional parameters for the OpenAI Responses Parse API.
 
         Returns:
             The response from OpenAI's responses.parse call.
         """
         return call_llm_and_track_usage(
-            posthog_distinct_id,
+            insights_distinct_id,
             self._client._ph_client,
             "openai",
-            posthog_trace_id,
-            posthog_properties,
-            posthog_privacy_mode,
-            posthog_groups,
+            insights_trace_id,
+            insights_properties,
+            insights_privacy_mode,
+            insights_groups,
             self._client.base_url,
             self._original.parse,
             **kwargs,
@@ -260,7 +260,7 @@ class WrappedResponses:
 
 
 class WrappedChat:
-    """Wrapper for OpenAI chat that tracks usage in PostHog."""
+    """Wrapper for OpenAI chat that tracks usage in Insights."""
 
     def __init__(self, client: OpenAI, original_chat):
         self._client = client
@@ -276,7 +276,7 @@ class WrappedChat:
 
 
 class WrappedCompletions:
-    """Wrapper for OpenAI chat completions that tracks usage in PostHog."""
+    """Wrapper for OpenAI chat completions that tracks usage in Insights."""
 
     def __init__(self, client: OpenAI, original_completions):
         self._client = client
@@ -288,34 +288,34 @@ class WrappedCompletions:
 
     def create(
         self,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_trace_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: bool = False,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_trace_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: bool = False,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs: Any,
     ):
-        if posthog_trace_id is None:
-            posthog_trace_id = str(uuid.uuid4())
+        if insights_trace_id is None:
+            insights_trace_id = str(uuid.uuid4())
 
         if kwargs.get("stream", False):
             return self._create_streaming(
-                posthog_distinct_id,
-                posthog_trace_id,
-                posthog_properties,
-                posthog_privacy_mode,
-                posthog_groups,
+                insights_distinct_id,
+                insights_trace_id,
+                insights_properties,
+                insights_privacy_mode,
+                insights_groups,
                 **kwargs,
             )
 
         return call_llm_and_track_usage(
-            posthog_distinct_id,
+            insights_distinct_id,
             self._client._ph_client,
             "openai",
-            posthog_trace_id,
-            posthog_properties,
-            posthog_privacy_mode,
-            posthog_groups,
+            insights_trace_id,
+            insights_properties,
+            insights_privacy_mode,
+            insights_groups,
             self._client.base_url,
             self._original.create,
             **kwargs,
@@ -323,11 +323,11 @@ class WrappedCompletions:
 
     def _create_streaming(
         self,
-        posthog_distinct_id: Optional[str],
-        posthog_trace_id: Optional[str],
-        posthog_properties: Optional[Dict[str, Any]],
-        posthog_privacy_mode: bool,
-        posthog_groups: Optional[Dict[str, Any]],
+        insights_distinct_id: Optional[str],
+        insights_trace_id: Optional[str],
+        insights_properties: Optional[Dict[str, Any]],
+        insights_privacy_mode: bool,
+        insights_groups: Optional[Dict[str, Any]],
         **kwargs: Any,
     ):
         start_time = time.time()
@@ -385,11 +385,11 @@ class WrappedCompletions:
                 )
 
                 self._capture_streaming_event(
-                    posthog_distinct_id,
-                    posthog_trace_id,
-                    posthog_properties,
-                    posthog_privacy_mode,
-                    posthog_groups,
+                    insights_distinct_id,
+                    insights_trace_id,
+                    insights_properties,
+                    insights_privacy_mode,
+                    insights_groups,
                     kwargs,
                     usage_stats,
                     latency,
@@ -403,11 +403,11 @@ class WrappedCompletions:
 
     def _capture_streaming_event(
         self,
-        posthog_distinct_id: Optional[str],
-        posthog_trace_id: Optional[str],
-        posthog_properties: Optional[Dict[str, Any]],
-        posthog_privacy_mode: bool,
-        posthog_groups: Optional[Dict[str, Any]],
+        insights_distinct_id: Optional[str],
+        insights_trace_id: Optional[str],
+        insights_properties: Optional[Dict[str, Any]],
+        insights_privacy_mode: bool,
+        insights_groups: Optional[Dict[str, Any]],
         kwargs: Dict[str, Any],
         usage_stats: TokenUsage,
         latency: float,
@@ -416,12 +416,12 @@ class WrappedCompletions:
         available_tool_calls: Optional[List[Dict[str, Any]]] = None,
         model_from_response: Optional[str] = None,
     ):
-        from posthog.ai.types import StreamingEventData
-        from posthog.ai.openai.openai_converter import (
+        from hanzo_insights.ai.types import StreamingEventData
+        from hanzo_insights.ai.openai.openai_converter import (
             format_openai_streaming_input,
             format_openai_streaming_output,
         )
-        from posthog.ai.utils import capture_streaming_event
+        from hanzo_insights.ai.utils import capture_streaming_event
 
         # Prepare standardized event data
         formatted_input = format_openai_streaming_input(kwargs, "chat")
@@ -439,11 +439,11 @@ class WrappedCompletions:
             formatted_output=format_openai_streaming_output(output, "chat", tool_calls),
             usage_stats=usage_stats,
             latency=latency,
-            distinct_id=posthog_distinct_id,
-            trace_id=posthog_trace_id,
-            properties=posthog_properties,
-            privacy_mode=posthog_privacy_mode,
-            groups=posthog_groups,
+            distinct_id=insights_distinct_id,
+            trace_id=insights_trace_id,
+            properties=insights_properties,
+            privacy_mode=insights_privacy_mode,
+            groups=insights_groups,
         )
 
         # Use the common capture function
@@ -451,7 +451,7 @@ class WrappedCompletions:
 
 
 class WrappedEmbeddings:
-    """Wrapper for OpenAI embeddings that tracks usage in PostHog."""
+    """Wrapper for OpenAI embeddings that tracks usage in Insights."""
 
     def __init__(self, client: OpenAI, original_embeddings):
         self._client = client
@@ -463,30 +463,30 @@ class WrappedEmbeddings:
 
     def create(
         self,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_trace_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: bool = False,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_trace_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: bool = False,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs: Any,
     ):
         """
-        Create an embedding using OpenAI's 'embeddings.create' method, but also track usage in PostHog.
+        Create an embedding using OpenAI's 'embeddings.create' method, but also track usage in Insights.
 
         Args:
-            posthog_distinct_id: Optional ID to associate with the usage event.
-            posthog_trace_id: Optional trace UUID for linking events.
-            posthog_properties: Optional dictionary of extra properties to include in the event.
-            posthog_privacy_mode: Whether to anonymize the input and output.
-            posthog_groups: Optional dictionary of groups to associate with the event.
+            insights_distinct_id: Optional ID to associate with the usage event.
+            insights_trace_id: Optional trace UUID for linking events.
+            insights_properties: Optional dictionary of extra properties to include in the event.
+            insights_privacy_mode: Whether to anonymize the input and output.
+            insights_groups: Optional dictionary of groups to associate with the event.
             **kwargs: Any additional parameters for the OpenAI Embeddings API.
 
         Returns:
             The response from OpenAI's embeddings.create call.
         """
 
-        if posthog_trace_id is None:
-            posthog_trace_id = str(uuid.uuid4())
+        if insights_trace_id is None:
+            insights_trace_id = str(uuid.uuid4())
 
         start_time = time.time()
         response = self._original.create(**kwargs)
@@ -508,34 +508,34 @@ class WrappedEmbeddings:
             "$ai_model": kwargs.get("model"),
             "$ai_input": with_privacy_mode(
                 self._client._ph_client,
-                posthog_privacy_mode,
+                insights_privacy_mode,
                 sanitize_openai_response(kwargs.get("input")),
             ),
             "$ai_http_status": 200,
             "$ai_input_tokens": usage_stats.get("prompt_tokens", 0),
             "$ai_latency": latency,
-            "$ai_trace_id": posthog_trace_id,
+            "$ai_trace_id": insights_trace_id,
             "$ai_base_url": str(self._client.base_url),
-            **(posthog_properties or {}),
+            **(insights_properties or {}),
         }
 
-        if posthog_distinct_id is None:
+        if insights_distinct_id is None:
             event_properties["$process_person_profile"] = False
 
         # Send capture event for embeddings
         if hasattr(self._client._ph_client, "capture"):
             self._client._ph_client.capture(
-                distinct_id=posthog_distinct_id or posthog_trace_id,
+                distinct_id=insights_distinct_id or insights_trace_id,
                 event="$ai_embedding",
                 properties=event_properties,
-                groups=posthog_groups,
+                groups=insights_groups,
             )
 
         return response
 
 
 class WrappedBeta:
-    """Wrapper for OpenAI beta features that tracks usage in PostHog."""
+    """Wrapper for OpenAI beta features that tracks usage in Insights."""
 
     def __init__(self, client: OpenAI, original_beta):
         self._client = client
@@ -551,7 +551,7 @@ class WrappedBeta:
 
 
 class WrappedBetaChat:
-    """Wrapper for OpenAI beta chat that tracks usage in PostHog."""
+    """Wrapper for OpenAI beta chat that tracks usage in Insights."""
 
     def __init__(self, client: OpenAI, original_beta_chat):
         self._client = client
@@ -567,7 +567,7 @@ class WrappedBetaChat:
 
 
 class WrappedBetaCompletions:
-    """Wrapper for OpenAI beta chat completions that tracks usage in PostHog."""
+    """Wrapper for OpenAI beta chat completions that tracks usage in Insights."""
 
     def __init__(self, client: OpenAI, original_beta_completions):
         self._client = client
@@ -579,21 +579,21 @@ class WrappedBetaCompletions:
 
     def parse(
         self,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_trace_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: bool = False,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_trace_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: bool = False,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs: Any,
     ):
         return call_llm_and_track_usage(
-            posthog_distinct_id,
+            insights_distinct_id,
             self._client._ph_client,
             "openai",
-            posthog_trace_id,
-            posthog_properties,
-            posthog_privacy_mode,
-            posthog_groups,
+            insights_trace_id,
+            insights_properties,
+            insights_privacy_mode,
+            insights_groups,
             self._client.base_url,
             self._original.parse,
             **kwargs,

hanzo_insights/ai/openai/openai_async.py

@@ -2,7 +2,7 @@ import time
 import uuid
 from typing import Any, Dict, List, Optional
 
-from posthog.ai.types import TokenUsage
+from hanzo_insights.ai.types import TokenUsage
 
 try:
     import openai
@@ -11,43 +11,43 @@ except ImportError:
         "Please install the OpenAI SDK to use this feature: 'pip install openai'"
     )
 
-from posthog import setup
-from posthog.ai.utils import (
+from hanzo_insights import setup
+from hanzo_insights.ai.utils import (
     call_llm_and_track_usage_async,
     extract_available_tool_calls,
     get_model_params,
     merge_usage_stats,
     with_privacy_mode,
 )
-from posthog.ai.openai.openai_converter import (
+from hanzo_insights.ai.openai.openai_converter import (
     extract_openai_usage_from_chunk,
     extract_openai_content_from_chunk,
     extract_openai_tool_calls_from_chunk,
     accumulate_openai_tool_calls,
     format_openai_streaming_output,
 )
-from posthog.ai.sanitization import sanitize_openai, sanitize_openai_response
-from posthog.client import Client as PostHogClient
+from hanzo_insights.ai.sanitization import sanitize_openai, sanitize_openai_response
+from hanzo_insights.client import Client as InsightsClient
 
 
 class AsyncOpenAI(openai.AsyncOpenAI):
     """
-    An async wrapper around the OpenAI SDK that automatically sends LLM usage events to PostHog.
+    An async wrapper around the OpenAI SDK that automatically sends LLM usage events to Insights.
     """
 
-    _ph_client: PostHogClient
+    _ph_client: InsightsClient
 
-    def __init__(self, posthog_client: Optional[PostHogClient] = None, **kwargs):
+    def __init__(self, insights_client: Optional[InsightsClient] = None, **kwargs):
         """
         Args:
             api_key: OpenAI API key.
-            posthog_client: If provided, events will be captured via this client instead
-                            of the global posthog.
+            insights_client: If provided, events will be captured via this client instead
+                            of the global hanzo_insights.
             **openai_config: Any additional keyword args to set on openai (e.g. organization="xxx").
         """
 
         super().__init__(**kwargs)
-        self._ph_client = posthog_client or setup()
+        self._ph_client = insights_client or setup()
 
         # Store original objects after parent initialization (only if they exist)
         self._original_chat = getattr(self, "chat", None)
@@ -70,7 +70,7 @@ class AsyncOpenAI(openai.AsyncOpenAI):
 
 
 class WrappedResponses:
-    """Async wrapper for OpenAI responses that tracks usage in PostHog."""
+    """Async wrapper for OpenAI responses that tracks usage in Insights."""
 
     def __init__(self, client: AsyncOpenAI, original_responses):
         self._client = client
@@ -83,34 +83,34 @@ class WrappedResponses:
 
     async def create(
         self,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_trace_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: bool = False,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_trace_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: bool = False,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs: Any,
     ):
-        if posthog_trace_id is None:
-            posthog_trace_id = str(uuid.uuid4())
+        if insights_trace_id is None:
+            insights_trace_id = str(uuid.uuid4())
 
         if kwargs.get("stream", False):
             return await self._create_streaming(
-                posthog_distinct_id,
-                posthog_trace_id,
-                posthog_properties,
-                posthog_privacy_mode,
-                posthog_groups,
+                insights_distinct_id,
+                insights_trace_id,
+                insights_properties,
+                insights_privacy_mode,
+                insights_groups,
                 **kwargs,
             )
 
         return await call_llm_and_track_usage_async(
-            posthog_distinct_id,
+            insights_distinct_id,
             self._client._ph_client,
             "openai",
-            posthog_trace_id,
-            posthog_properties,
-            posthog_privacy_mode,
-            posthog_groups,
+            insights_trace_id,
+            insights_properties,
+            insights_privacy_mode,
+            insights_groups,
             self._client.base_url,
             self._original.create,
             **kwargs,
@@ -118,11 +118,11 @@ class WrappedResponses:
 
     async def _create_streaming(
         self,
-        posthog_distinct_id: Optional[str],
-        posthog_trace_id: Optional[str],
-        posthog_properties: Optional[Dict[str, Any]],
-        posthog_privacy_mode: bool,
-        posthog_groups: Optional[Dict[str, Any]],
+        insights_distinct_id: Optional[str],
+        insights_trace_id: Optional[str],
+        insights_properties: Optional[Dict[str, Any]],
+        insights_privacy_mode: bool,
+        insights_groups: Optional[Dict[str, Any]],
         **kwargs: Any,
     ):
         start_time = time.time()
@@ -165,11 +165,11 @@ class WrappedResponses:
                 output = final_content
 
                 await self._capture_streaming_event(
-                    posthog_distinct_id,
-                    posthog_trace_id,
-                    posthog_properties,
-                    posthog_privacy_mode,
-                    posthog_groups,
+                    insights_distinct_id,
+                    insights_trace_id,
+                    insights_properties,
+                    insights_privacy_mode,
+                    insights_groups,
                     kwargs,
                     usage_stats,
                     latency,
@@ -182,11 +182,11 @@ class WrappedResponses:
 
     async def _capture_streaming_event(
         self,
-        posthog_distinct_id: Optional[str],
-        posthog_trace_id: Optional[str],
-        posthog_properties: Optional[Dict[str, Any]],
-        posthog_privacy_mode: bool,
-        posthog_groups: Optional[Dict[str, Any]],
+        insights_distinct_id: Optional[str],
+        insights_trace_id: Optional[str],
+        insights_properties: Optional[Dict[str, Any]],
+        insights_privacy_mode: bool,
+        insights_groups: Optional[Dict[str, Any]],
         kwargs: Dict[str, Any],
         usage_stats: TokenUsage,
         latency: float,
@@ -194,8 +194,8 @@ class WrappedResponses:
         available_tool_calls: Optional[List[Dict[str, Any]]] = None,
         model_from_response: Optional[str] = None,
     ):
-        if posthog_trace_id is None:
-            posthog_trace_id = str(uuid.uuid4())
+        if insights_trace_id is None:
+            insights_trace_id = str(uuid.uuid4())
 
         # Use model from kwargs, fallback to model from response
         model = kwargs.get("model") or model_from_response or "unknown"
@@ -206,12 +206,12 @@ class WrappedResponses:
             "$ai_model_parameters": get_model_params(kwargs),
             "$ai_input": with_privacy_mode(
                 self._client._ph_client,
-                posthog_privacy_mode,
+                insights_privacy_mode,
                 sanitize_openai_response(kwargs.get("input")),
             ),
             "$ai_output_choices": with_privacy_mode(
                 self._client._ph_client,
-                posthog_privacy_mode,
+                insights_privacy_mode,
                 format_openai_streaming_output(output, "responses"),
             ),
             "$ai_http_status": 200,
@@ -222,9 +222,9 @@ class WrappedResponses:
             ),
             "$ai_reasoning_tokens": usage_stats.get("reasoning_tokens", 0),
             "$ai_latency": latency,
-            "$ai_trace_id": posthog_trace_id,
+            "$ai_trace_id": insights_trace_id,
             "$ai_base_url": str(self._client.base_url),
-            **(posthog_properties or {}),
+            **(insights_properties or {}),
         }
 
         # Add web search count if present
@@ -239,48 +239,48 @@ class WrappedResponses:
         if available_tool_calls:
             event_properties["$ai_tools"] = available_tool_calls
 
-        if posthog_distinct_id is None:
+        if insights_distinct_id is None:
             event_properties["$process_person_profile"] = False
 
         if hasattr(self._client._ph_client, "capture"):
             self._client._ph_client.capture(
-                distinct_id=posthog_distinct_id or posthog_trace_id,
+                distinct_id=insights_distinct_id or insights_trace_id,
                 event="$ai_generation",
                 properties=event_properties,
-                groups=posthog_groups,
+                groups=insights_groups,
             )
 
     async def parse(
         self,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_trace_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: bool = False,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_trace_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: bool = False,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs: Any,
     ):
         """
-        Parse structured output using OpenAI's 'responses.parse' method, but also track usage in PostHog.
+        Parse structured output using OpenAI's 'responses.parse' method, but also track usage in Insights.
 
         Args:
-            posthog_distinct_id: Optional ID to associate with the usage event.
-            posthog_trace_id: Optional trace UUID for linking events.
-            posthog_properties: Optional dictionary of extra properties to include in the event.
-            posthog_privacy_mode: Whether to anonymize the input and output.
-            posthog_groups: Optional dictionary of groups to associate with the event.
+            insights_distinct_id: Optional ID to associate with the usage event.
+            insights_trace_id: Optional trace UUID for linking events.
+            insights_properties: Optional dictionary of extra properties to include in the event.
+            insights_privacy_mode: Whether to anonymize the input and output.
+            insights_groups: Optional dictionary of groups to associate with the event.
             **kwargs: Any additional parameters for the OpenAI Responses Parse API.
 
         Returns:
             The response from OpenAI's responses.parse call.
         """
         return await call_llm_and_track_usage_async(
-            posthog_distinct_id,
+            insights_distinct_id,
             self._client._ph_client,
             "openai",
-            posthog_trace_id,
-            posthog_properties,
-            posthog_privacy_mode,
-            posthog_groups,
+            insights_trace_id,
+            insights_properties,
+            insights_privacy_mode,
+            insights_groups,
             self._client.base_url,
             self._original.parse,
             **kwargs,
@@ -288,7 +288,7 @@ class WrappedResponses:
 
 
 class WrappedChat:
-    """Async wrapper for OpenAI chat that tracks usage in PostHog."""
+    """Async wrapper for OpenAI chat that tracks usage in Insights."""
 
     def __init__(self, client: AsyncOpenAI, original_chat):
         self._client = client
@@ -304,7 +304,7 @@ class WrappedChat:
 
 
 class WrappedCompletions:
-    """Async wrapper for OpenAI chat completions that tracks usage in PostHog."""
+    """Async wrapper for OpenAI chat completions that tracks usage in Insights."""
 
     def __init__(self, client: AsyncOpenAI, original_completions):
         self._client = client
@@ -316,35 +316,35 @@ class WrappedCompletions:
 
     async def create(
         self,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_trace_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: bool = False,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_trace_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: bool = False,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs: Any,
     ):
-        if posthog_trace_id is None:
-            posthog_trace_id = str(uuid.uuid4())
+        if insights_trace_id is None:
+            insights_trace_id = str(uuid.uuid4())
 
         # If streaming, handle streaming specifically
         if kwargs.get("stream", False):
             return await self._create_streaming(
-                posthog_distinct_id,
-                posthog_trace_id,
-                posthog_properties,
-                posthog_privacy_mode,
-                posthog_groups,
+                insights_distinct_id,
+                insights_trace_id,
+                insights_properties,
+                insights_privacy_mode,
+                insights_groups,
                 **kwargs,
             )
 
         response = await call_llm_and_track_usage_async(
-            posthog_distinct_id,
+            insights_distinct_id,
             self._client._ph_client,
             "openai",
-            posthog_trace_id,
-            posthog_properties,
-            posthog_privacy_mode,
-            posthog_groups,
+            insights_trace_id,
+            insights_properties,
+            insights_privacy_mode,
+            insights_groups,
             self._client.base_url,
             self._original.create,
             **kwargs,
@@ -353,11 +353,11 @@ class WrappedCompletions:
 
     async def _create_streaming(
         self,
-        posthog_distinct_id: Optional[str],
-        posthog_trace_id: Optional[str],
-        posthog_properties: Optional[Dict[str, Any]],
-        posthog_privacy_mode: bool,
-        posthog_groups: Optional[Dict[str, Any]],
+        insights_distinct_id: Optional[str],
+        insights_trace_id: Optional[str],
+        insights_properties: Optional[Dict[str, Any]],
+        insights_privacy_mode: bool,
+        insights_groups: Optional[Dict[str, Any]],
         **kwargs: Any,
     ):
         start_time = time.time()
@@ -414,11 +414,11 @@ class WrappedCompletions:
                 )
 
                 await self._capture_streaming_event(
-                    posthog_distinct_id,
-                    posthog_trace_id,
-                    posthog_properties,
-                    posthog_privacy_mode,
-                    posthog_groups,
+                    insights_distinct_id,
+                    insights_trace_id,
+                    insights_properties,
+                    insights_privacy_mode,
+                    insights_groups,
                     kwargs,
                     usage_stats,
                     latency,
@@ -432,11 +432,11 @@ class WrappedCompletions:
 
     async def _capture_streaming_event(
         self,
-        posthog_distinct_id: Optional[str],
-        posthog_trace_id: Optional[str],
-        posthog_properties: Optional[Dict[str, Any]],
-        posthog_privacy_mode: bool,
-        posthog_groups: Optional[Dict[str, Any]],
+        insights_distinct_id: Optional[str],
+        insights_trace_id: Optional[str],
+        insights_properties: Optional[Dict[str, Any]],
+        insights_privacy_mode: bool,
+        insights_groups: Optional[Dict[str, Any]],
         kwargs: Dict[str, Any],
         usage_stats: TokenUsage,
         latency: float,
@@ -445,8 +445,8 @@ class WrappedCompletions:
         available_tool_calls: Optional[List[Dict[str, Any]]] = None,
         model_from_response: Optional[str] = None,
     ):
-        if posthog_trace_id is None:
-            posthog_trace_id = str(uuid.uuid4())
+        if insights_trace_id is None:
+            insights_trace_id = str(uuid.uuid4())
 
         # Use model from kwargs, fallback to model from response
         model = kwargs.get("model") or model_from_response or "unknown"
@@ -457,12 +457,12 @@ class WrappedCompletions:
             "$ai_model_parameters": get_model_params(kwargs),
             "$ai_input": with_privacy_mode(
                 self._client._ph_client,
-                posthog_privacy_mode,
+                insights_privacy_mode,
                 sanitize_openai(kwargs.get("messages")),
             ),
             "$ai_output_choices": with_privacy_mode(
                 self._client._ph_client,
-                posthog_privacy_mode,
+                insights_privacy_mode,
                 format_openai_streaming_output(output, "chat", tool_calls),
             ),
             "$ai_http_status": 200,
@@ -473,9 +473,9 @@ class WrappedCompletions:
             ),
             "$ai_reasoning_tokens": usage_stats.get("reasoning_tokens", 0),
             "$ai_latency": latency,
-            "$ai_trace_id": posthog_trace_id,
+            "$ai_trace_id": insights_trace_id,
             "$ai_base_url": str(self._client.base_url),
-            **(posthog_properties or {}),
+            **(insights_properties or {}),
         }
 
         # Add web search count if present
@@ -491,20 +491,20 @@ class WrappedCompletions:
         if available_tool_calls:
             event_properties["$ai_tools"] = available_tool_calls
 
-        if posthog_distinct_id is None:
+        if insights_distinct_id is None:
             event_properties["$process_person_profile"] = False
 
         if hasattr(self._client._ph_client, "capture"):
             self._client._ph_client.capture(
-                distinct_id=posthog_distinct_id or posthog_trace_id,
+                distinct_id=insights_distinct_id or insights_trace_id,
                 event="$ai_generation",
                 properties=event_properties,
-                groups=posthog_groups,
+                groups=insights_groups,
             )
 
 
 class WrappedEmbeddings:
-    """Async wrapper for OpenAI embeddings that tracks usage in PostHog."""
+    """Async wrapper for OpenAI embeddings that tracks usage in Insights."""
 
     def __init__(self, client: AsyncOpenAI, original_embeddings):
         self._client = client
@@ -517,30 +517,30 @@ class WrappedEmbeddings:
 
     async def create(
         self,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_trace_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: bool = False,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_trace_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: bool = False,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs: Any,
     ):
         """
-        Create an embedding using OpenAI's 'embeddings.create' method, but also track usage in PostHog.
+        Create an embedding using OpenAI's 'embeddings.create' method, but also track usage in Insights.
 
         Args:
-            posthog_distinct_id: Optional ID to associate with the usage event.
-            posthog_trace_id: Optional trace UUID for linking events.
-            posthog_properties: Optional dictionary of extra properties to include in the event.
-            posthog_privacy_mode: Whether to anonymize the input and output.
-            posthog_groups: Optional dictionary of groups to associate with the event.
+            insights_distinct_id: Optional ID to associate with the usage event.
+            insights_trace_id: Optional trace UUID for linking events.
+            insights_properties: Optional dictionary of extra properties to include in the event.
+            insights_privacy_mode: Whether to anonymize the input and output.
+            insights_groups: Optional dictionary of groups to associate with the event.
             **kwargs: Any additional parameters for the OpenAI Embeddings API.
 
         Returns:
             The response from OpenAI's embeddings.create call.
         """
 
-        if posthog_trace_id is None:
-            posthog_trace_id = str(uuid.uuid4())
+        if insights_trace_id is None:
+            insights_trace_id = str(uuid.uuid4())
 
         start_time = time.time()
         response = await self._original.create(**kwargs)
@@ -563,34 +563,34 @@ class WrappedEmbeddings:
             "$ai_model": kwargs.get("model"),
             "$ai_input": with_privacy_mode(
                 self._client._ph_client,
-                posthog_privacy_mode,
+                insights_privacy_mode,
                 sanitize_openai_response(kwargs.get("input")),
             ),
             "$ai_http_status": 200,
             "$ai_input_tokens": usage_stats.get("input_tokens", 0),
             "$ai_latency": latency,
-            "$ai_trace_id": posthog_trace_id,
+            "$ai_trace_id": insights_trace_id,
             "$ai_base_url": str(self._client.base_url),
-            **(posthog_properties or {}),
+            **(insights_properties or {}),
         }
 
-        if posthog_distinct_id is None:
+        if insights_distinct_id is None:
             event_properties["$process_person_profile"] = False
 
         # Send capture event for embeddings
         if hasattr(self._client._ph_client, "capture"):
             self._client._ph_client.capture(
-                distinct_id=posthog_distinct_id or posthog_trace_id,
+                distinct_id=insights_distinct_id or insights_trace_id,
                 event="$ai_embedding",
                 properties=event_properties,
-                groups=posthog_groups,
+                groups=insights_groups,
             )
 
         return response
 
 
 class WrappedBeta:
-    """Async wrapper for OpenAI beta features that tracks usage in PostHog."""
+    """Async wrapper for OpenAI beta features that tracks usage in Insights."""
 
     def __init__(self, client: AsyncOpenAI, original_beta):
         self._client = client
@@ -607,7 +607,7 @@ class WrappedBeta:
 
 
 class WrappedBetaChat:
-    """Async wrapper for OpenAI beta chat that tracks usage in PostHog."""
+    """Async wrapper for OpenAI beta chat that tracks usage in Insights."""
 
     def __init__(self, client: AsyncOpenAI, original_beta_chat):
         self._client = client
@@ -624,7 +624,7 @@ class WrappedBetaChat:
 
 
 class WrappedBetaCompletions:
-    """Async wrapper for OpenAI beta chat completions that tracks usage in PostHog."""
+    """Async wrapper for OpenAI beta chat completions that tracks usage in Insights."""
 
     def __init__(self, client: AsyncOpenAI, original_beta_completions):
         self._client = client
@@ -637,21 +637,21 @@ class WrappedBetaCompletions:
 
     async def parse(
         self,
-        posthog_distinct_id: Optional[str] = None,
-        posthog_trace_id: Optional[str] = None,
-        posthog_properties: Optional[Dict[str, Any]] = None,
-        posthog_privacy_mode: bool = False,
-        posthog_groups: Optional[Dict[str, Any]] = None,
+        insights_distinct_id: Optional[str] = None,
+        insights_trace_id: Optional[str] = None,
+        insights_properties: Optional[Dict[str, Any]] = None,
+        insights_privacy_mode: bool = False,
+        insights_groups: Optional[Dict[str, Any]] = None,
         **kwargs: Any,
     ):
         return await call_llm_and_track_usage_async(
-            posthog_distinct_id,
+            insights_distinct_id,
             self._client._ph_client,
             "openai",
-            posthog_trace_id,
-            posthog_properties,
-            posthog_privacy_mode,
-            posthog_groups,
+            insights_trace_id,
+            insights_properties,
+            insights_privacy_mode,
+            insights_groups,
             self._client.base_url,
             self._original.parse,
             **kwargs,

hanzo_insights/ai/openai/openai_converter.py

@@ -2,13 +2,13 @@
 OpenAI-specific conversion utilities.
 
 This module handles the conversion of OpenAI API responses and inputs
-into standardized formats for PostHog tracking. It supports both
+into standardized formats for Insights tracking. It supports both
 Chat Completions API and Responses API formats.
 """
 
 from typing import Any, Dict, List, Optional
 
-from posthog.ai.types import (
+from hanzo_insights.ai.types import (
     FormattedContentItem,
     FormattedFunctionCall,
     FormattedImageContent,
@@ -16,6 +16,7 @@ from posthog.ai.types import (
     FormattedTextContent,
     TokenUsage,
 )
+from hanzo_insights.ai.utils import serialize_raw_usage
 
 
 def format_openai_response(response: Any) -> List[FormattedMessage]:
@@ -429,6 +430,12 @@ def extract_openai_usage_from_response(response: Any) -> TokenUsage:
     if web_search_count > 0:
         result["web_search_count"] = web_search_count
 
+    # Capture raw usage metadata for backend processing
+    # Serialize to dict here in the converter (not in utils)
+    serialized = serialize_raw_usage(response.usage)
+    if serialized:
+        result["raw_usage"] = serialized
+
     return result
 
 
@@ -482,6 +489,12 @@ def extract_openai_usage_from_chunk(
                 chunk.usage.completion_tokens_details.reasoning_tokens
             )
 
+        # Capture raw usage metadata for backend processing
+        # Serialize to dict here in the converter (not in utils)
+        serialized = serialize_raw_usage(chunk.usage)
+        if serialized:
+            usage["raw_usage"] = serialized
+
     elif provider_type == "responses":
         # For Responses API, usage is only in chunk.response.usage for completed events
         if hasattr(chunk, "type") and chunk.type == "response.completed":
@@ -516,6 +529,12 @@ def extract_openai_usage_from_chunk(
                     if web_search_count > 0:
                         usage["web_search_count"] = web_search_count
 
+                # Capture raw usage metadata for backend processing
+                # Serialize to dict here in the converter (not in utils)
+                serialized = serialize_raw_usage(response_usage)
+                if serialized:
+                    usage["raw_usage"] = serialized
+
     return usage
 
 
@@ -734,8 +753,8 @@ def format_openai_streaming_input(
         api_type: Either "chat" or "responses"
 
     Returns:
-        Formatted input ready for PostHog tracking
+        Formatted input ready for Insights tracking
     """
-    from posthog.ai.utils import merge_system_prompt
+    from hanzo_insights.ai.utils import merge_system_prompt
 
     return merge_system_prompt(kwargs, "openai")

hanzo_insights/ai/openai/openai_providers.py

@@ -5,39 +5,39 @@ except ImportError:
         "Please install the Open AI SDK to use this feature: 'pip install openai'"
     )
 
-from posthog.ai.openai.openai import (
+from hanzo_insights.ai.openai.openai import (
     WrappedBeta,
     WrappedChat,
     WrappedEmbeddings,
     WrappedResponses,
 )
-from posthog.ai.openai.openai_async import WrappedBeta as AsyncWrappedBeta
-from posthog.ai.openai.openai_async import WrappedChat as AsyncWrappedChat
-from posthog.ai.openai.openai_async import WrappedEmbeddings as AsyncWrappedEmbeddings
-from posthog.ai.openai.openai_async import WrappedResponses as AsyncWrappedResponses
+from hanzo_insights.ai.openai.openai_async import WrappedBeta as AsyncWrappedBeta
+from hanzo_insights.ai.openai.openai_async import WrappedChat as AsyncWrappedChat
+from hanzo_insights.ai.openai.openai_async import WrappedEmbeddings as AsyncWrappedEmbeddings
+from hanzo_insights.ai.openai.openai_async import WrappedResponses as AsyncWrappedResponses
 from typing import Optional
 
-from posthog.client import Client as PostHogClient
-from posthog import setup
+from hanzo_insights.client import Client as InsightsClient
+from hanzo_insights import setup
 
 
 class AzureOpenAI(openai.AzureOpenAI):
     """
-    A wrapper around the Azure OpenAI SDK that automatically sends LLM usage events to PostHog.
+    A wrapper around the Azure OpenAI SDK that automatically sends LLM usage events to Insights.
     """
 
-    _ph_client: PostHogClient
+    _ph_client: InsightsClient
 
-    def __init__(self, posthog_client: Optional[PostHogClient] = None, **kwargs):
+    def __init__(self, insights_client: Optional[InsightsClient] = None, **kwargs):
         """
         Args:
             api_key: Azure OpenAI API key.
-            posthog_client: If provided, events will be captured via this client instead
-                            of the global posthog.
+            insights_client: If provided, events will be captured via this client instead
+                            of the global hanzo_insights.
             **openai_config: Any additional keyword args to set on Azure OpenAI (e.g. azure_endpoint="xxx").
         """
         super().__init__(**kwargs)
-        self._ph_client = posthog_client or setup()
+        self._ph_client = insights_client or setup()
 
         # Store original objects after parent initialization (only if they exist)
         self._original_chat = getattr(self, "chat", None)
@@ -61,21 +61,21 @@ class AzureOpenAI(openai.AzureOpenAI):
 
 class AsyncAzureOpenAI(openai.AsyncAzureOpenAI):
     """
-    An async wrapper around the Azure OpenAI SDK that automatically sends LLM usage events to PostHog.
+    An async wrapper around the Azure OpenAI SDK that automatically sends LLM usage events to Insights.
     """
 
-    _ph_client: PostHogClient
+    _ph_client: InsightsClient
 
-    def __init__(self, posthog_client: Optional[PostHogClient] = None, **kwargs):
+    def __init__(self, insights_client: Optional[InsightsClient] = None, **kwargs):
         """
         Args:
             api_key: Azure OpenAI API key.
-            posthog_client: If provided, events will be captured via this client instead
-                            of the global posthog.
+            insights_client: If provided, events will be captured via this client instead
+                            of the global hanzo_insights.
             **openai_config: Any additional keyword args to set on Azure OpenAI (e.g. azure_endpoint="xxx").
         """
         super().__init__(**kwargs)
-        self._ph_client = posthog_client or setup()
+        self._ph_client = insights_client or setup()
 
         # Store original objects after parent initialization (only if they exist)
         self._original_chat = getattr(self, "chat", None)

hanzo_insights/ai/openai_agents/__init__.py

@@ -0,0 +1,76 @@
+from __future__ import annotations
+
+from typing import TYPE_CHECKING, Any, Callable, Dict, Optional, Union
+
+if TYPE_CHECKING:
+    from agents.tracing import Trace
+
+    from hanzo_insights.client import Client
+
+try:
+    import agents  # noqa: F401
+except ImportError:
+    raise ModuleNotFoundError(
+        "Please install the OpenAI Agents SDK to use this feature: 'pip install openai-agents'"
+    )
+
+from hanzo_insights.ai.openai_agents.processor import InsightsTracingProcessor
+
+__all__ = ["InsightsTracingProcessor", "instrument"]
+
+
+def instrument(
+    client: Optional[Client] = None,
+    distinct_id: Optional[Union[str, Callable[[Trace], Optional[str]]]] = None,
+    privacy_mode: bool = False,
+    groups: Optional[Dict[str, Any]] = None,
+    properties: Optional[Dict[str, Any]] = None,
+) -> InsightsTracingProcessor:
+    """
+    One-liner to instrument OpenAI Agents SDK with Hanzo Insights tracing.
+
+    This registers an InsightsTracingProcessor with the OpenAI Agents SDK,
+    automatically capturing traces, spans, and LLM generations.
+
+    Args:
+        client: Optional Insights client instance. If not provided, uses the default client.
+        distinct_id: Optional distinct ID to associate with all traces.
+            Can also be a callable that takes a trace and returns a distinct ID.
+        privacy_mode: If True, redacts input/output content from events.
+        groups: Optional Insights groups to associate with events.
+        properties: Optional additional properties to include with all events.
+
+    Returns:
+        InsightsTracingProcessor: The registered processor instance.
+
+    Example:
+        ```python
+        from hanzo_insights.ai.openai_agents import instrument
+
+        # Simple setup
+        instrument(distinct_id="user@example.com")
+
+        # With custom properties
+        instrument(
+            distinct_id="user@example.com",
+            privacy_mode=True,
+            properties={"environment": "production"}
+        )
+
+        # Now run agents as normal - traces automatically sent to Insights
+        from agents import Agent, Runner
+        agent = Agent(name="Assistant", instructions="You are helpful.")
+        result = Runner.run_sync(agent, "Hello!")
+        ```
+    """
+    from agents.tracing import add_trace_processor
+
+    processor = InsightsTracingProcessor(
+        client=client,
+        distinct_id=distinct_id,
+        privacy_mode=privacy_mode,
+        groups=groups,
+        properties=properties,
+    )
+    add_trace_processor(processor)
+    return processor

hanzo_insights/ai/openai_agents/processor.py

@@ -0,0 +1,863 @@
+import json
+import logging
+import time
+from datetime import datetime
+from typing import Any, Callable, Dict, Optional, Union
+
+from agents.tracing import Span, Trace
+from agents.tracing.processor_interface import TracingProcessor
+from agents.tracing.span_data import (
+    AgentSpanData,
+    CustomSpanData,
+    FunctionSpanData,
+    GenerationSpanData,
+    GuardrailSpanData,
+    HandoffSpanData,
+    MCPListToolsSpanData,
+    ResponseSpanData,
+    SpeechGroupSpanData,
+    SpeechSpanData,
+    TranscriptionSpanData,
+)
+
+from hanzo_insights import setup
+from hanzo_insights.client import Client
+
+log = logging.getLogger("hanzo_insights")
+
+
+def _ensure_serializable(obj: Any) -> Any:
+    """Ensure an object is JSON-serializable, converting to str as fallback.
+
+    Returns the original object if it's already serializable (dict, list, str,
+    int, etc.), or str(obj) for non-serializable types so that downstream
+    json.dumps() calls won't fail.
+    """
+    if obj is None:
+        return None
+    try:
+        json.dumps(obj)
+        return obj
+    except (TypeError, ValueError):
+        return str(obj)
+
+
+def _parse_iso_timestamp(iso_str: Optional[str]) -> Optional[float]:
+    """Parse ISO timestamp to Unix timestamp."""
+    if not iso_str:
+        return None
+    try:
+        dt = datetime.fromisoformat(iso_str.replace("Z", "+00:00"))
+        return dt.timestamp()
+    except (ValueError, AttributeError):
+        return None
+
+
+class InsightsTracingProcessor(TracingProcessor):
+    """
+    A tracing processor that sends OpenAI Agents SDK traces to Hanzo Insights.
+
+    This processor implements the TracingProcessor interface from the OpenAI Agents SDK
+    and maps agent traces, spans, and generations to Insights LLM analytics events.
+
+    Example:
+        ```python
+        from agents import Agent, Runner
+        from agents.tracing import add_trace_processor
+        from hanzo_insights.ai.openai_agents import InsightsTracingProcessor
+
+        # Create and register the processor
+        processor = InsightsTracingProcessor(
+            distinct_id="user@example.com",
+            privacy_mode=False,
+        )
+        add_trace_processor(processor)
+
+        # Run agents as normal - traces automatically sent to Insights
+        agent = Agent(name="Assistant", instructions="You are helpful.")
+        result = Runner.run_sync(agent, "Hello!")
+        ```
+    """
+
+    def __init__(
+        self,
+        client: Optional[Client] = None,
+        distinct_id: Optional[Union[str, Callable[[Trace], Optional[str]]]] = None,
+        privacy_mode: bool = False,
+        groups: Optional[Dict[str, Any]] = None,
+        properties: Optional[Dict[str, Any]] = None,
+    ):
+        """
+        Initialize the Insights tracing processor.
+
+        Args:
+            client: Optional Insights client instance. If not provided, uses the default client.
+            distinct_id: Either a string distinct ID or a callable that takes a Trace
+                and returns a distinct ID. If not provided, uses the trace_id.
+            privacy_mode: If True, redacts input/output content from events.
+            groups: Optional Insights groups to associate with all events.
+            properties: Optional additional properties to include with all events.
+        """
+        self._client = client or setup()
+        self._distinct_id = distinct_id
+        self._privacy_mode = privacy_mode
+        self._groups = groups or {}
+        self._properties = properties or {}
+
+        # Track span start times for latency calculation
+        self._span_start_times: Dict[str, float] = {}
+
+        # Track trace metadata for associating with spans
+        self._trace_metadata: Dict[str, Dict[str, Any]] = {}
+
+        # Max entries to prevent unbounded growth if on_span_end/on_trace_end
+        # is never called (e.g., due to an exception in the Agents SDK).
+        self._max_tracked_entries = 10000
+
+    def _get_distinct_id(self, trace: Optional[Trace]) -> Optional[str]:
+        """Resolve the distinct ID for a trace.
+
+        Returns the user-provided distinct ID (string or callable result),
+        or None if no user-provided ID is available. Callers should treat
+        None as a signal to use a fallback ID in personless mode.
+        """
+        if callable(self._distinct_id):
+            if trace:
+                result = self._distinct_id(trace)
+                if result:
+                    return str(result)
+            return None
+        elif self._distinct_id:
+            return str(self._distinct_id)
+        return None
+
+    def _with_privacy_mode(self, value: Any) -> Any:
+        """Apply privacy mode redaction if enabled."""
+        if self._privacy_mode or (
+            hasattr(self._client, "privacy_mode") and self._client.privacy_mode
+        ):
+            return None
+        return value
+
+    def _evict_stale_entries(self) -> None:
+        """Evict oldest entries if dicts exceed max size to prevent unbounded growth."""
+        if len(self._span_start_times) > self._max_tracked_entries:
+            # Remove oldest entries by start time
+            sorted_spans = sorted(self._span_start_times.items(), key=lambda x: x[1])
+            for span_id, _ in sorted_spans[: len(sorted_spans) // 2]:
+                del self._span_start_times[span_id]
+            log.debug(
+                "Evicted stale span start times (exceeded %d entries)",
+                self._max_tracked_entries,
+            )
+
+        if len(self._trace_metadata) > self._max_tracked_entries:
+            # Remove half the entries (oldest inserted via dict ordering in Python 3.7+)
+            keys = list(self._trace_metadata.keys())
+            for key in keys[: len(keys) // 2]:
+                del self._trace_metadata[key]
+            log.debug(
+                "Evicted stale trace metadata (exceeded %d entries)",
+                self._max_tracked_entries,
+            )
+
+    def _get_group_id(self, trace_id: str) -> Optional[str]:
+        """Get the group_id for a trace from stored metadata."""
+        if trace_id in self._trace_metadata:
+            return self._trace_metadata[trace_id].get("group_id")
+        return None
+
+    def _capture_event(
+        self,
+        event: str,
+        properties: Dict[str, Any],
+        distinct_id: Optional[str] = None,
+    ) -> None:
+        """Capture an event to Insights with error handling.
+
+        Args:
+            distinct_id: The resolved distinct ID. When the user didn't provide
+                one, callers should pass ``user_distinct_id or fallback_id``
+                (matching the langchain/openai pattern) and separately set
+                ``$process_person_profile`` in properties.
+        """
+        try:
+            if not hasattr(self._client, "capture") or not callable(
+                self._client.capture
+            ):
+                return
+
+            final_properties = {
+                **properties,
+                **self._properties,
+            }
+
+            self._client.capture(
+                distinct_id=distinct_id or "unknown",
+                event=event,
+                properties=final_properties,
+                groups=self._groups,
+            )
+        except Exception as e:
+            log.debug(f"Failed to capture Insights event: {e}")
+
+    def on_trace_start(self, trace: Trace) -> None:
+        """Called when a new trace begins. Stores metadata for spans; the $ai_trace event is emitted in on_trace_end."""
+        try:
+            self._evict_stale_entries()
+            trace_id = trace.trace_id
+            trace_name = trace.name
+            group_id = getattr(trace, "group_id", None)
+            metadata = getattr(trace, "metadata", None)
+
+            distinct_id = self._get_distinct_id(trace)
+
+            # Store trace metadata for later (used by spans and on_trace_end)
+            self._trace_metadata[trace_id] = {
+                "name": trace_name,
+                "group_id": group_id,
+                "metadata": metadata,
+                "distinct_id": distinct_id,
+                "start_time": time.time(),
+            }
+        except Exception as e:
+            log.debug(f"Error in on_trace_start: {e}")
+
+    def on_trace_end(self, trace: Trace) -> None:
+        """Called when a trace completes. Emits the $ai_trace event with full metadata."""
+        try:
+            trace_id = trace.trace_id
+
+            # Pop stored metadata (also cleans up)
+            trace_info = self._trace_metadata.pop(trace_id, {})
+            trace_name = trace_info.get("name") or trace.name
+            group_id = trace_info.get("group_id") or getattr(trace, "group_id", None)
+            metadata = trace_info.get("metadata") or getattr(trace, "metadata", None)
+            distinct_id = trace_info.get("distinct_id") or self._get_distinct_id(trace)
+
+            # Calculate trace-level latency
+            start_time = trace_info.get("start_time")
+            latency = (time.time() - start_time) if start_time else None
+
+            properties = {
+                "$ai_trace_id": trace_id,
+                "$ai_trace_name": trace_name,
+                "$ai_provider": "openai",
+                "$ai_framework": "openai-agents",
+            }
+
+            if latency is not None:
+                properties["$ai_latency"] = latency
+
+            # Include group_id for linking related traces (e.g., conversation threads)
+            if group_id:
+                properties["$ai_group_id"] = group_id
+
+            # Include trace metadata if present
+            if metadata:
+                properties["$ai_trace_metadata"] = _ensure_serializable(metadata)
+
+            if distinct_id is None:
+                properties["$process_person_profile"] = False
+
+            self._capture_event(
+                event="$ai_trace",
+                distinct_id=distinct_id or trace_id,
+                properties=properties,
+            )
+        except Exception as e:
+            log.debug(f"Error in on_trace_end: {e}")
+
+    def on_span_start(self, span: Span[Any]) -> None:
+        """Called when a new span begins."""
+        try:
+            self._evict_stale_entries()
+            span_id = span.span_id
+            self._span_start_times[span_id] = time.time()
+        except Exception as e:
+            log.debug(f"Error in on_span_start: {e}")
+
+    def on_span_end(self, span: Span[Any]) -> None:
+        """Called when a span completes."""
+        try:
+            span_id = span.span_id
+            trace_id = span.trace_id
+            parent_id = span.parent_id
+            span_data = span.span_data
+
+            # Calculate latency
+            start_time = self._span_start_times.pop(span_id, None)
+            if start_time:
+                latency = time.time() - start_time
+            else:
+                # Fall back to parsing timestamps
+                started = _parse_iso_timestamp(span.started_at)
+                ended = _parse_iso_timestamp(span.ended_at)
+                latency = (ended - started) if (started and ended) else 0
+
+            # Get user-provided distinct ID from trace metadata (resolved at trace start).
+            # None means no user-provided ID — use trace_id as fallback in personless mode,
+            # matching the langchain/openai pattern: `distinct_id or trace_id`.
+            trace_info = self._trace_metadata.get(trace_id, {})
+            distinct_id = trace_info.get("distinct_id") or self._get_distinct_id(None)
+
+            # Get group_id from trace metadata for linking
+            group_id = self._get_group_id(trace_id)
+
+            # Get error info if present
+            error_info = span.error
+            error_properties = {}
+            if error_info:
+                if isinstance(error_info, dict):
+                    error_message = error_info.get("message", str(error_info))
+                    error_type_raw = error_info.get("type", "")
+                else:
+                    error_message = str(error_info)
+                    error_type_raw = ""
+
+                # Categorize error type for cross-provider filtering/alerting
+                error_type = "unknown"
+                if (
+                    "ModelBehaviorError" in error_type_raw
+                    or "ModelBehaviorError" in error_message
+                ):
+                    error_type = "model_behavior_error"
+                elif "UserError" in error_type_raw or "UserError" in error_message:
+                    error_type = "user_error"
+                elif (
+                    "InputGuardrailTripwireTriggered" in error_type_raw
+                    or "InputGuardrailTripwireTriggered" in error_message
+                ):
+                    error_type = "input_guardrail_triggered"
+                elif (
+                    "OutputGuardrailTripwireTriggered" in error_type_raw
+                    or "OutputGuardrailTripwireTriggered" in error_message
+                ):
+                    error_type = "output_guardrail_triggered"
+                elif (
+                    "MaxTurnsExceeded" in error_type_raw
+                    or "MaxTurnsExceeded" in error_message
+                ):
+                    error_type = "max_turns_exceeded"
+
+                error_properties = {
+                    "$ai_is_error": True,
+                    "$ai_error": error_message,
+                    "$ai_error_type": error_type,
+                }
+
+            # Personless mode: no user-provided distinct_id, fallback to trace_id
+            if distinct_id is None:
+                error_properties["$process_person_profile"] = False
+                distinct_id = trace_id
+
+            # Dispatch based on span data type
+            if isinstance(span_data, GenerationSpanData):
+                self._handle_generation_span(
+                    span_data,
+                    trace_id,
+                    span_id,
+                    parent_id,
+                    latency,
+                    distinct_id,
+                    group_id,
+                    error_properties,
+                )
+            elif isinstance(span_data, FunctionSpanData):
+                self._handle_function_span(
+                    span_data,
+                    trace_id,
+                    span_id,
+                    parent_id,
+                    latency,
+                    distinct_id,
+                    group_id,
+                    error_properties,
+                )
+            elif isinstance(span_data, AgentSpanData):
+                self._handle_agent_span(
+                    span_data,
+                    trace_id,
+                    span_id,
+                    parent_id,
+                    latency,
+                    distinct_id,
+                    group_id,
+                    error_properties,
+                )
+            elif isinstance(span_data, HandoffSpanData):
+                self._handle_handoff_span(
+                    span_data,
+                    trace_id,
+                    span_id,
+                    parent_id,
+                    latency,
+                    distinct_id,
+                    group_id,
+                    error_properties,
+                )
+            elif isinstance(span_data, GuardrailSpanData):
+                self._handle_guardrail_span(
+                    span_data,
+                    trace_id,
+                    span_id,
+                    parent_id,
+                    latency,
+                    distinct_id,
+                    group_id,
+                    error_properties,
+                )
+            elif isinstance(span_data, ResponseSpanData):
+                self._handle_response_span(
+                    span_data,
+                    trace_id,
+                    span_id,
+                    parent_id,
+                    latency,
+                    distinct_id,
+                    group_id,
+                    error_properties,
+                )
+            elif isinstance(span_data, CustomSpanData):
+                self._handle_custom_span(
+                    span_data,
+                    trace_id,
+                    span_id,
+                    parent_id,
+                    latency,
+                    distinct_id,
+                    group_id,
+                    error_properties,
+                )
+            elif isinstance(
+                span_data, (TranscriptionSpanData, SpeechSpanData, SpeechGroupSpanData)
+            ):
+                self._handle_audio_span(
+                    span_data,
+                    trace_id,
+                    span_id,
+                    parent_id,
+                    latency,
+                    distinct_id,
+                    group_id,
+                    error_properties,
+                )
+            elif isinstance(span_data, MCPListToolsSpanData):
+                self._handle_mcp_span(
+                    span_data,
+                    trace_id,
+                    span_id,
+                    parent_id,
+                    latency,
+                    distinct_id,
+                    group_id,
+                    error_properties,
+                )
+            else:
+                # Unknown span type - capture as generic span
+                self._handle_generic_span(
+                    span_data,
+                    trace_id,
+                    span_id,
+                    parent_id,
+                    latency,
+                    distinct_id,
+                    group_id,
+                    error_properties,
+                )
+
+        except Exception as e:
+            log.debug(f"Error in on_span_end: {e}")
+
+    def _base_properties(
+        self,
+        trace_id: str,
+        span_id: str,
+        parent_id: Optional[str],
+        latency: float,
+        group_id: Optional[str],
+        error_properties: Dict[str, Any],
+    ) -> Dict[str, Any]:
+        """Build the base properties dict shared by all span handlers."""
+        properties = {
+            "$ai_trace_id": trace_id,
+            "$ai_span_id": span_id,
+            "$ai_parent_id": parent_id,
+            "$ai_provider": "openai",
+            "$ai_framework": "openai-agents",
+            "$ai_latency": latency,
+            **error_properties,
+        }
+        if group_id:
+            properties["$ai_group_id"] = group_id
+        return properties
+
+    def _handle_generation_span(
+        self,
+        span_data: GenerationSpanData,
+        trace_id: str,
+        span_id: str,
+        parent_id: Optional[str],
+        latency: float,
+        distinct_id: str,
+        group_id: Optional[str],
+        error_properties: Dict[str, Any],
+    ) -> None:
+        """Handle LLM generation spans - maps to $ai_generation event."""
+        # Extract token usage
+        usage = span_data.usage or {}
+        input_tokens = usage.get("input_tokens") or usage.get("prompt_tokens") or 0
+        output_tokens = (
+            usage.get("output_tokens") or usage.get("completion_tokens") or 0
+        )
+
+        # Extract model config parameters
+        model_config = span_data.model_config or {}
+        model_params = {}
+        for param in [
+            "temperature",
+            "max_tokens",
+            "top_p",
+            "frequency_penalty",
+            "presence_penalty",
+        ]:
+            if param in model_config:
+                model_params[param] = model_config[param]
+
+        properties = {
+            **self._base_properties(
+                trace_id, span_id, parent_id, latency, group_id, error_properties
+            ),
+            "$ai_model": span_data.model,
+            "$ai_model_parameters": model_params if model_params else None,
+            "$ai_input": self._with_privacy_mode(_ensure_serializable(span_data.input)),
+            "$ai_output_choices": self._with_privacy_mode(
+                _ensure_serializable(span_data.output)
+            ),
+            "$ai_input_tokens": input_tokens,
+            "$ai_output_tokens": output_tokens,
+            "$ai_total_tokens": (input_tokens or 0) + (output_tokens or 0),
+        }
+
+        # Add optional token fields if present
+        if usage.get("reasoning_tokens"):
+            properties["$ai_reasoning_tokens"] = usage["reasoning_tokens"]
+        if usage.get("cache_read_input_tokens"):
+            properties["$ai_cache_read_input_tokens"] = usage["cache_read_input_tokens"]
+        if usage.get("cache_creation_input_tokens"):
+            properties["$ai_cache_creation_input_tokens"] = usage[
+                "cache_creation_input_tokens"
+            ]
+
+        self._capture_event("$ai_generation", properties, distinct_id)
+
+    def _handle_function_span(
+        self,
+        span_data: FunctionSpanData,
+        trace_id: str,
+        span_id: str,
+        parent_id: Optional[str],
+        latency: float,
+        distinct_id: str,
+        group_id: Optional[str],
+        error_properties: Dict[str, Any],
+    ) -> None:
+        """Handle function/tool call spans - maps to $ai_span event."""
+        properties = {
+            **self._base_properties(
+                trace_id, span_id, parent_id, latency, group_id, error_properties
+            ),
+            "$ai_span_name": span_data.name,
+            "$ai_span_type": "tool",
+            "$ai_input_state": self._with_privacy_mode(
+                _ensure_serializable(span_data.input)
+            ),
+            "$ai_output_state": self._with_privacy_mode(
+                _ensure_serializable(span_data.output)
+            ),
+        }
+
+        if span_data.mcp_data:
+            properties["$ai_mcp_data"] = _ensure_serializable(span_data.mcp_data)
+
+        self._capture_event("$ai_span", properties, distinct_id)
+
+    def _handle_agent_span(
+        self,
+        span_data: AgentSpanData,
+        trace_id: str,
+        span_id: str,
+        parent_id: Optional[str],
+        latency: float,
+        distinct_id: str,
+        group_id: Optional[str],
+        error_properties: Dict[str, Any],
+    ) -> None:
+        """Handle agent execution spans - maps to $ai_span event."""
+        properties = {
+            **self._base_properties(
+                trace_id, span_id, parent_id, latency, group_id, error_properties
+            ),
+            "$ai_span_name": span_data.name,
+            "$ai_span_type": "agent",
+        }
+
+        if span_data.handoffs:
+            properties["$ai_agent_handoffs"] = span_data.handoffs
+        if span_data.tools:
+            properties["$ai_agent_tools"] = span_data.tools
+        if span_data.output_type:
+            properties["$ai_agent_output_type"] = span_data.output_type
+
+        self._capture_event("$ai_span", properties, distinct_id)
+
+    def _handle_handoff_span(
+        self,
+        span_data: HandoffSpanData,
+        trace_id: str,
+        span_id: str,
+        parent_id: Optional[str],
+        latency: float,
+        distinct_id: str,
+        group_id: Optional[str],
+        error_properties: Dict[str, Any],
+    ) -> None:
+        """Handle agent handoff spans - maps to $ai_span event."""
+        properties = {
+            **self._base_properties(
+                trace_id, span_id, parent_id, latency, group_id, error_properties
+            ),
+            "$ai_span_name": f"{span_data.from_agent} -> {span_data.to_agent}",
+            "$ai_span_type": "handoff",
+            "$ai_handoff_from_agent": span_data.from_agent,
+            "$ai_handoff_to_agent": span_data.to_agent,
+        }
+
+        self._capture_event("$ai_span", properties, distinct_id)
+
+    def _handle_guardrail_span(
+        self,
+        span_data: GuardrailSpanData,
+        trace_id: str,
+        span_id: str,
+        parent_id: Optional[str],
+        latency: float,
+        distinct_id: str,
+        group_id: Optional[str],
+        error_properties: Dict[str, Any],
+    ) -> None:
+        """Handle guardrail execution spans - maps to $ai_span event."""
+        properties = {
+            **self._base_properties(
+                trace_id, span_id, parent_id, latency, group_id, error_properties
+            ),
+            "$ai_span_name": span_data.name,
+            "$ai_span_type": "guardrail",
+            "$ai_guardrail_triggered": span_data.triggered,
+        }
+
+        self._capture_event("$ai_span", properties, distinct_id)
+
+    def _handle_response_span(
+        self,
+        span_data: ResponseSpanData,
+        trace_id: str,
+        span_id: str,
+        parent_id: Optional[str],
+        latency: float,
+        distinct_id: str,
+        group_id: Optional[str],
+        error_properties: Dict[str, Any],
+    ) -> None:
+        """Handle OpenAI Response API spans - maps to $ai_generation event."""
+        response = span_data.response
+        response_id = response.id if response else None
+
+        # Try to extract usage from response
+        usage = getattr(response, "usage", None) if response else None
+        input_tokens = 0
+        output_tokens = 0
+        if usage:
+            input_tokens = getattr(usage, "input_tokens", 0) or 0
+            output_tokens = getattr(usage, "output_tokens", 0) or 0
+
+        # Try to extract model from response
+        model = getattr(response, "model", None) if response else None
+
+        properties = {
+            **self._base_properties(
+                trace_id, span_id, parent_id, latency, group_id, error_properties
+            ),
+            "$ai_model": model,
+            "$ai_response_id": response_id,
+            "$ai_input": self._with_privacy_mode(_ensure_serializable(span_data.input)),
+            "$ai_input_tokens": input_tokens,
+            "$ai_output_tokens": output_tokens,
+            "$ai_total_tokens": input_tokens + output_tokens,
+        }
+
+        # Extract output content from response
+        if response:
+            output_items = getattr(response, "output", None)
+            if output_items:
+                properties["$ai_output_choices"] = self._with_privacy_mode(
+                    _ensure_serializable(output_items)
+                )
+
+        self._capture_event("$ai_generation", properties, distinct_id)
+
+    def _handle_custom_span(
+        self,
+        span_data: CustomSpanData,
+        trace_id: str,
+        span_id: str,
+        parent_id: Optional[str],
+        latency: float,
+        distinct_id: str,
+        group_id: Optional[str],
+        error_properties: Dict[str, Any],
+    ) -> None:
+        """Handle custom user-defined spans - maps to $ai_span event."""
+        properties = {
+            **self._base_properties(
+                trace_id, span_id, parent_id, latency, group_id, error_properties
+            ),
+            "$ai_span_name": span_data.name,
+            "$ai_span_type": "custom",
+            "$ai_custom_data": self._with_privacy_mode(
+                _ensure_serializable(span_data.data)
+            ),
+        }
+
+        self._capture_event("$ai_span", properties, distinct_id)
+
+    def _handle_audio_span(
+        self,
+        span_data: Union[TranscriptionSpanData, SpeechSpanData, SpeechGroupSpanData],
+        trace_id: str,
+        span_id: str,
+        parent_id: Optional[str],
+        latency: float,
+        distinct_id: str,
+        group_id: Optional[str],
+        error_properties: Dict[str, Any],
+    ) -> None:
+        """Handle audio-related spans (transcription, speech) - maps to $ai_span event."""
+        span_type = span_data.type  # "transcription", "speech", or "speech_group"
+
+        properties = {
+            **self._base_properties(
+                trace_id, span_id, parent_id, latency, group_id, error_properties
+            ),
+            "$ai_span_name": span_type,
+            "$ai_span_type": span_type,
+        }
+
+        # Add model info if available
+        if hasattr(span_data, "model") and span_data.model:
+            properties["$ai_model"] = span_data.model
+
+        # Add model config if available (pass-through property)
+        if hasattr(span_data, "model_config") and span_data.model_config:
+            properties["model_config"] = _ensure_serializable(span_data.model_config)
+
+        # Add time to first audio byte for speech spans (pass-through property)
+        if hasattr(span_data, "first_content_at") and span_data.first_content_at:
+            properties["first_content_at"] = span_data.first_content_at
+
+        # Add audio format info (pass-through properties)
+        if hasattr(span_data, "input_format"):
+            properties["audio_input_format"] = span_data.input_format
+        if hasattr(span_data, "output_format"):
+            properties["audio_output_format"] = span_data.output_format
+
+        # Add text input for TTS
+        if (
+            hasattr(span_data, "input")
+            and span_data.input
+            and isinstance(span_data.input, str)
+        ):
+            properties["$ai_input"] = self._with_privacy_mode(span_data.input)
+
+        # Don't include audio data (base64) - just metadata
+        if hasattr(span_data, "output") and isinstance(span_data.output, str):
+            # For transcription, output is the text
+            properties["$ai_output_state"] = self._with_privacy_mode(span_data.output)
+
+        self._capture_event("$ai_span", properties, distinct_id)
+
+    def _handle_mcp_span(
+        self,
+        span_data: MCPListToolsSpanData,
+        trace_id: str,
+        span_id: str,
+        parent_id: Optional[str],
+        latency: float,
+        distinct_id: str,
+        group_id: Optional[str],
+        error_properties: Dict[str, Any],
+    ) -> None:
+        """Handle MCP (Model Context Protocol) spans - maps to $ai_span event."""
+        properties = {
+            **self._base_properties(
+                trace_id, span_id, parent_id, latency, group_id, error_properties
+            ),
+            "$ai_span_name": f"mcp:{span_data.server}",
+            "$ai_span_type": "mcp_tools",
+            "$ai_mcp_server": span_data.server,
+            "$ai_mcp_tools": span_data.result,
+        }
+
+        self._capture_event("$ai_span", properties, distinct_id)
+
+    def _handle_generic_span(
+        self,
+        span_data: Any,
+        trace_id: str,
+        span_id: str,
+        parent_id: Optional[str],
+        latency: float,
+        distinct_id: str,
+        group_id: Optional[str],
+        error_properties: Dict[str, Any],
+    ) -> None:
+        """Handle unknown span types - maps to $ai_span event."""
+        span_type = getattr(span_data, "type", "unknown")
+
+        properties = {
+            **self._base_properties(
+                trace_id, span_id, parent_id, latency, group_id, error_properties
+            ),
+            "$ai_span_name": span_type,
+            "$ai_span_type": span_type,
+        }
+
+        # Try to export span data
+        if hasattr(span_data, "export"):
+            try:
+                exported = span_data.export()
+                properties["$ai_span_data"] = _ensure_serializable(exported)
+            except Exception:
+                pass
+
+        self._capture_event("$ai_span", properties, distinct_id)
+
+    def shutdown(self) -> None:
+        """Clean up resources when the application stops."""
+        try:
+            self._span_start_times.clear()
+            self._trace_metadata.clear()
+
+            # Flush the Insights client if possible
+            if hasattr(self._client, "flush") and callable(self._client.flush):
+                self._client.flush()
+        except Exception as e:
+            log.debug(f"Error in shutdown: {e}")
+
+    def force_flush(self) -> None:
+        """Force immediate processing of any queued events."""
+        try:
+            if hasattr(self._client, "flush") and callable(self._client.flush):
+                self._client.flush()
+        except Exception as e:
+            log.debug(f"Error in force_flush: {e}")

hanzo_insights/ai/prompts.py

@@ -0,0 +1,329 @@
+"""
+Prompt management for Hanzo Insights AI SDK.
+
+Fetch and compile LLM prompts from Insights with caching and fallback support.
+"""
+
+import logging
+import re
+import time
+import urllib.parse
+from typing import Any, Dict, Optional, Union
+
+from hanzo_insights.request import USER_AGENT, _get_session
+from hanzo_insights.utils import remove_trailing_slash
+
+log = logging.getLogger("hanzo_insights")
+
+APP_ENDPOINT = "https://us.insights.hanzo.ai"
+DEFAULT_CACHE_TTL_SECONDS = 300  # 5 minutes
+
+PromptVariables = Dict[str, Union[str, int, float, bool]]
+PromptCacheKey = tuple[str, Optional[int]]
+
+
+class CachedPrompt:
+    """Cached prompt with metadata."""
+
+    def __init__(self, prompt: str, fetched_at: float):
+        self.prompt = prompt
+        self.fetched_at = fetched_at
+
+
+def _cache_key(name: str, version: Optional[int]) -> PromptCacheKey:
+    """Build a cache key for latest or versioned prompt fetches."""
+    return (name, version)
+
+
+def _prompt_reference(name: str, version: Optional[int]) -> str:
+    """Format a prompt reference for logs and errors."""
+    label = f'prompt "{name}"'
+    if version is not None:
+        return f"{label} version {version}"
+    return label
+
+
+def _is_prompt_api_response(data: Any) -> bool:
+    """Check if the response is a valid prompt API response."""
+    return (
+        isinstance(data, dict)
+        and "prompt" in data
+        and isinstance(data.get("prompt"), str)
+    )
+
+
+class Prompts:
+    """
+    Fetch and compile LLM prompts from Insights.
+
+    Can be initialized with a Insights client or with direct options.
+
+    Examples:
+        ```python
+        from hanzo_insights import Insights
+        from hanzo_insights.ai.prompts import Prompts
+
+        # With Insights client
+        client = Insights('phc_xxx', host='https://us.insights.hanzo.ai', personal_api_key='phx_xxx')
+        prompts = Prompts(client)
+
+        # Or with direct options (no Insights client needed)
+        prompts = Prompts(
+            personal_api_key='phx_xxx',
+            project_api_key='phc_xxx',
+            host='https://us.insights.hanzo.ai',
+        )
+
+        # Fetch with caching and fallback
+        template = prompts.get('support-system-prompt', fallback='You are a helpful assistant.')
+
+        # Fetch a specific published version
+        prompt_v1 = prompts.get('support-system-prompt', version=1)
+
+        # Compile with variables
+        system_prompt = prompts.compile(template, {
+            'company': 'Acme Corp',
+            'tier': 'premium',
+        })
+        ```
+    """
+
+    def __init__(
+        self,
+        client: Optional[Any] = None,
+        *,
+        personal_api_key: Optional[str] = None,
+        project_api_key: Optional[str] = None,
+        host: Optional[str] = None,
+        default_cache_ttl_seconds: Optional[int] = None,
+    ):
+        """
+        Initialize Prompts.
+
+        Args:
+            client: Insights client instance (optional if personal_api_key provided)
+            personal_api_key: Direct personal API key (optional if client provided)
+            project_api_key: Direct project API key (optional if client provided)
+            host: Insights host (defaults to app endpoint)
+            default_cache_ttl_seconds: Default cache TTL (defaults to 300)
+        """
+        self._default_cache_ttl_seconds = (
+            default_cache_ttl_seconds or DEFAULT_CACHE_TTL_SECONDS
+        )
+        self._cache: Dict[PromptCacheKey, CachedPrompt] = {}
+
+        if client is not None:
+            self._personal_api_key = getattr(client, "personal_api_key", None) or ""
+            self._project_api_key = getattr(client, "api_key", None) or ""
+            self._host = remove_trailing_slash(
+                getattr(client, "raw_host", None) or APP_ENDPOINT
+            )
+        else:
+            self._personal_api_key = personal_api_key or ""
+            self._project_api_key = project_api_key or ""
+            self._host = remove_trailing_slash(host or APP_ENDPOINT)
+
+    def get(
+        self,
+        name: str,
+        *,
+        cache_ttl_seconds: Optional[int] = None,
+        fallback: Optional[str] = None,
+        version: Optional[int] = None,
+    ) -> str:
+        """
+        Fetch a prompt by name from the Insights API.
+
+        Caching behavior:
+        1. If cache is fresh, return cached value
+        2. If fetch fails and cache exists (stale), return stale cache with warning
+        3. If fetch fails and fallback provided, return fallback with warning
+        4. If fetch fails with no cache/fallback, raise exception
+
+        Args:
+            name: The name of the prompt to fetch
+            cache_ttl_seconds: Cache TTL in seconds (defaults to instance default)
+            fallback: Fallback prompt to use if fetch fails and no cache available
+            version: Specific prompt version to fetch. If None, fetches the latest
+                version
+
+        Returns:
+            The prompt string
+
+        Raises:
+            Exception: If the prompt cannot be fetched and no fallback is available
+        """
+        ttl = (
+            cache_ttl_seconds
+            if cache_ttl_seconds is not None
+            else self._default_cache_ttl_seconds
+        )
+        cache_key = _cache_key(name, version)
+
+        # Check cache first
+        cached = self._cache.get(cache_key)
+        now = time.time()
+
+        if cached is not None:
+            is_fresh = (now - cached.fetched_at) < ttl
+
+            if is_fresh:
+                return cached.prompt
+
+        # Try to fetch from API
+        try:
+            prompt = self._fetch_prompt_from_api(name, version)
+            fetched_at = time.time()
+
+            # Update cache
+            self._cache[cache_key] = CachedPrompt(prompt=prompt, fetched_at=fetched_at)
+
+            return prompt
+
+        except Exception as error:
+            prompt_reference = _prompt_reference(name, version)
+            # Fallback order:
+            # 1. Return stale cache (with warning)
+            if cached is not None:
+                log.warning(
+                    "[Insights Prompts] Failed to fetch %s, using stale cache: %s",
+                    prompt_reference,
+                    error,
+                )
+                return cached.prompt
+
+            # 2. Return fallback (with warning)
+            if fallback is not None:
+                log.warning(
+                    "[Insights Prompts] Failed to fetch %s, using fallback: %s",
+                    prompt_reference,
+                    error,
+                )
+                return fallback
+
+            # 3. Raise error
+            raise
+
+    def compile(self, prompt: str, variables: PromptVariables) -> str:
+        """
+        Replace {{variableName}} placeholders with values.
+
+        Unmatched variables are left unchanged.
+        Supports variable names with hyphens and dots (e.g., user-id, company.name).
+
+        Args:
+            prompt: The prompt template string
+            variables: Object containing variable values
+
+        Returns:
+            The compiled prompt string
+        """
+
+        def replace_variable(match: re.Match) -> str:
+            variable_name = match.group(1)
+
+            if variable_name in variables:
+                return str(variables[variable_name])
+
+            return match.group(0)
+
+        return re.sub(r"\{\{([\w.-]+)\}\}", replace_variable, prompt)
+
+    def clear_cache(
+        self, name: Optional[str] = None, *, version: Optional[int] = None
+    ) -> None:
+        """
+        Clear cached prompts.
+
+        Args:
+            name: Specific prompt name to clear. If None, clears all cached prompts.
+            version: Specific prompt version to clear. Requires name.
+        """
+        if version is not None and name is None:
+            raise ValueError("'version' requires 'name' to be provided")
+
+        if name is None:
+            self._cache.clear()
+            return
+
+        if version is not None:
+            self._cache.pop(_cache_key(name, version), None)
+            return
+
+        keys_to_clear = [key for key in self._cache if key[0] == name]
+        for key in keys_to_clear:
+            self._cache.pop(key, None)
+
+    def _fetch_prompt_from_api(self, name: str, version: Optional[int] = None) -> str:
+        """
+        Fetch prompt from Insights API.
+
+        Endpoint:
+            {host}/api/environments/@current/llm_prompts/name/{encoded_name}/
+            ?token={encoded_project_api_key}[&version={version}]
+        Auth: Bearer {personal_api_key}
+
+        Args:
+            name: The name of the prompt to fetch
+            version: Specific prompt version to fetch. If None, fetches the latest
+
+        Returns:
+            The prompt string
+
+        Raises:
+            Exception: If the prompt cannot be fetched
+        """
+        if not self._personal_api_key:
+            raise Exception(
+                "[Insights Prompts] personal_api_key is required to fetch prompts. "
+                "Please provide it when initializing the Prompts instance."
+            )
+        if not self._project_api_key:
+            raise Exception(
+                "[Insights Prompts] project_api_key is required to fetch prompts. "
+                "Please provide it when initializing the Prompts instance."
+            )
+
+        encoded_name = urllib.parse.quote(name, safe="")
+        query_params: Dict[str, Union[str, int]] = {"token": self._project_api_key}
+        if version is not None:
+            query_params["version"] = version
+        encoded_query = urllib.parse.urlencode(query_params)
+        url = f"{self._host}/api/environments/@current/llm_prompts/name/{encoded_name}/?{encoded_query}"
+        prompt_reference = _prompt_reference(name, version)
+        prompt_label = prompt_reference[:1].upper() + prompt_reference[1:]
+
+        headers = {
+            "Authorization": f"Bearer {self._personal_api_key}",
+            "User-Agent": USER_AGENT,
+        }
+
+        response = _get_session().get(url, headers=headers, timeout=10)
+
+        if not response.ok:
+            if response.status_code == 404:
+                raise Exception(f"[Insights Prompts] {prompt_label} not found")
+
+            if response.status_code == 403:
+                raise Exception(
+                    f"[Insights Prompts] Access denied for {prompt_reference}. "
+                    "Check that your personal_api_key has the correct permissions and the LLM prompts feature is enabled."
+                )
+
+            raise Exception(
+                f"[Insights Prompts] Failed to fetch {prompt_label}: HTTP {response.status_code}"
+            )
+
+        try:
+            data = response.json()
+        except Exception:
+            raise Exception(
+                f"[Insights Prompts] Invalid response format for {prompt_label}"
+            )
+
+        if not _is_prompt_api_response(data):
+            raise Exception(
+                f"[Insights Prompts] Invalid response format for {prompt_label}"
+            )
+
+        return data["prompt"]

hanzo_insights/ai/sanitization.py

@@ -83,6 +83,12 @@ def sanitize_openai_image(item: Any) -> Any:
     if not isinstance(item, dict):
         return item
 
+    if item.get("type") == "input_image" and isinstance(item.get("image_url"), str):
+        return {
+            **item,
+            "image_url": redact_base64_data_url(item["image_url"]),
+        }
+
     if (
         item.get("type") == "image_url"
         and isinstance(item.get("image_url"), dict)

hanzo_insights/ai/types.py

@@ -1,5 +1,5 @@
 """
-Common type definitions for PostHog AI SDK.
+Common type definitions for Insights AI SDK.
 
 These types are used for formatting messages and responses across different AI providers
 (Anthropic, OpenAI, Gemini, etc.) to ensure consistency in tracking and data structure.
@@ -41,10 +41,10 @@ FormattedContentItem = Union[
 
 class FormattedMessage(TypedDict):
     """
-    Standardized message format for PostHog tracking.
+    Standardized message format for Insights tracking.
 
     Used across all providers to ensure consistent message structure
-    when sending events to PostHog.
+    when sending events to Insights.
     """
 
     role: str
@@ -64,6 +64,7 @@ class TokenUsage(TypedDict, total=False):
     cache_creation_input_tokens: Optional[int]
     reasoning_tokens: Optional[int]
     web_search_count: Optional[int]
+    raw_usage: Optional[Any]  # Raw provider usage metadata for backend processing
 
 
 class ProviderResponse(TypedDict, total=False):

hanzo_insights/ai/utils.py

@@ -0,0 +1,757 @@
+import time
+import uuid
+from typing import Any, Callable, Dict, List, Optional, cast
+
+from hanzo_insights import get_tags, identify_context, new_context, tag, contexts
+from hanzo_insights.ai.sanitization import (
+    sanitize_anthropic,
+    sanitize_gemini,
+    sanitize_langchain,
+    sanitize_openai,
+)
+from hanzo_insights.ai.types import FormattedMessage, StreamingEventData, TokenUsage
+from hanzo_insights.client import Client as InsightsClient
+
+
+_TOKEN_PROPERTY_KEYS = frozenset(
+    {
+        "$ai_input_tokens",
+        "$ai_output_tokens",
+        "$ai_cache_read_input_tokens",
+        "$ai_cache_creation_input_tokens",
+        "$ai_total_tokens",
+        "$ai_reasoning_tokens",
+    }
+)
+
+
+def _get_tokens_source(
+    sdk_tags: Dict[str, Any], insights_properties: Optional[Dict[str, Any]]
+) -> str:
+    if insights_properties and any(
+        key in insights_properties for key in _TOKEN_PROPERTY_KEYS
+    ):
+        return "passthrough"
+    return "sdk"
+
+
+def serialize_raw_usage(raw_usage: Any) -> Optional[Dict[str, Any]]:
+    """
+    Convert raw provider usage objects to JSON-serializable dicts.
+
+    Handles Pydantic models (OpenAI/Anthropic) and protobuf-like objects (Gemini)
+    with a fallback chain to ensure we never pass unserializable objects to Insights.
+
+    Args:
+        raw_usage: Raw usage object from provider SDK
+
+    Returns:
+        Plain dict or None if conversion fails
+    """
+    if raw_usage is None:
+        return None
+
+    # Already a dict
+    if isinstance(raw_usage, dict):
+        return raw_usage
+
+    # Try Pydantic model_dump() (OpenAI/Anthropic)
+    if hasattr(raw_usage, "model_dump") and callable(raw_usage.model_dump):
+        try:
+            return raw_usage.model_dump()
+        except Exception:
+            pass
+
+    # Try to_dict() (some protobuf objects)
+    if hasattr(raw_usage, "to_dict") and callable(raw_usage.to_dict):
+        try:
+            return raw_usage.to_dict()
+        except Exception:
+            pass
+
+    # Try __dict__ / vars() for simple objects
+    try:
+        return vars(raw_usage)
+    except Exception:
+        pass
+
+    # Last resort: convert to string representation
+    # This ensures we always return something rather than failing
+    try:
+        return {"_raw": str(raw_usage)}
+    except Exception:
+        return None
+
+
+def merge_usage_stats(
+    target: TokenUsage, source: TokenUsage, mode: str = "incremental"
+) -> None:
+    """
+    Merge streaming usage statistics into target dict, handling None values.
+
+    Supports two modes:
+    - "incremental": Add source values to target (for APIs that report new tokens)
+    - "cumulative": Replace target with source values (for APIs that report totals)
+
+    Args:
+        target: Dictionary to update with usage stats
+        source: TokenUsage that may contain None values
+        mode: Either "incremental" or "cumulative"
+    """
+    if mode == "incremental":
+        # Add new values to existing totals
+        source_input = source.get("input_tokens")
+        if source_input is not None:
+            current = target.get("input_tokens") or 0
+            target["input_tokens"] = current + source_input
+
+        source_output = source.get("output_tokens")
+        if source_output is not None:
+            current = target.get("output_tokens") or 0
+            target["output_tokens"] = current + source_output
+
+        source_cache_read = source.get("cache_read_input_tokens")
+        if source_cache_read is not None:
+            current = target.get("cache_read_input_tokens") or 0
+            target["cache_read_input_tokens"] = current + source_cache_read
+
+        source_cache_creation = source.get("cache_creation_input_tokens")
+        if source_cache_creation is not None:
+            current = target.get("cache_creation_input_tokens") or 0
+            target["cache_creation_input_tokens"] = current + source_cache_creation
+
+        source_reasoning = source.get("reasoning_tokens")
+        if source_reasoning is not None:
+            current = target.get("reasoning_tokens") or 0
+            target["reasoning_tokens"] = current + source_reasoning
+
+        source_web_search = source.get("web_search_count")
+        if source_web_search is not None:
+            current = target.get("web_search_count") or 0
+            target["web_search_count"] = max(current, source_web_search)
+
+        # Merge raw_usage to avoid losing data from earlier events
+        # For Anthropic streaming: message_start has input tokens, message_delta has output
+        # Note: raw_usage is already serialized by converters, so it's a dict
+        source_raw_usage = source.get("raw_usage")
+        if source_raw_usage is not None and isinstance(source_raw_usage, dict):
+            current_raw_value = target.get("raw_usage")
+            current_raw: Dict[str, Any] = (
+                current_raw_value if isinstance(current_raw_value, dict) else {}
+            )
+            target["raw_usage"] = {**current_raw, **source_raw_usage}
+
+    elif mode == "cumulative":
+        # Replace with latest values (already cumulative)
+        if source.get("input_tokens") is not None:
+            target["input_tokens"] = source["input_tokens"]
+        if source.get("output_tokens") is not None:
+            target["output_tokens"] = source["output_tokens"]
+        if source.get("cache_read_input_tokens") is not None:
+            target["cache_read_input_tokens"] = source["cache_read_input_tokens"]
+        if source.get("cache_creation_input_tokens") is not None:
+            target["cache_creation_input_tokens"] = source[
+                "cache_creation_input_tokens"
+            ]
+        if source.get("reasoning_tokens") is not None:
+            target["reasoning_tokens"] = source["reasoning_tokens"]
+        if source.get("web_search_count") is not None:
+            target["web_search_count"] = source["web_search_count"]
+        # Note: raw_usage is already serialized by converters, so it's a dict
+        if source.get("raw_usage") is not None:
+            target["raw_usage"] = source["raw_usage"]
+
+    else:
+        raise ValueError(f"Invalid mode: {mode}. Must be 'incremental' or 'cumulative'")
+
+
+def get_model_params(kwargs: Dict[str, Any]) -> Dict[str, Any]:
+    """
+    Extracts model parameters from the kwargs dictionary.
+    """
+    model_params = {}
+    for param in [
+        "temperature",
+        "max_tokens",  # Deprecated field
+        "max_completion_tokens",
+        "top_p",
+        "frequency_penalty",
+        "presence_penalty",
+        "n",
+        "stop",
+        "stream",  # OpenAI-specific field
+        "streaming",  # Anthropic-specific field
+    ]:
+        if param in kwargs and kwargs[param] is not None:
+            model_params[param] = kwargs[param]
+    return model_params
+
+
+def get_usage(response, provider: str) -> TokenUsage:
+    """
+    Extract usage statistics from response based on provider.
+    Delegates to provider-specific converter functions.
+    """
+    if provider == "anthropic":
+        from hanzo_insights.ai.anthropic.anthropic_converter import (
+            extract_anthropic_usage_from_response,
+        )
+
+        return extract_anthropic_usage_from_response(response)
+    elif provider == "openai":
+        from hanzo_insights.ai.openai.openai_converter import (
+            extract_openai_usage_from_response,
+        )
+
+        return extract_openai_usage_from_response(response)
+    elif provider == "gemini":
+        from hanzo_insights.ai.gemini.gemini_converter import (
+            extract_gemini_usage_from_response,
+        )
+
+        return extract_gemini_usage_from_response(response)
+
+    return TokenUsage(input_tokens=0, output_tokens=0)
+
+
+def format_response(response, provider: str):
+    """
+    Format a regular (non-streaming) response.
+    """
+    if provider == "anthropic":
+        from hanzo_insights.ai.anthropic.anthropic_converter import format_anthropic_response
+
+        return format_anthropic_response(response)
+    elif provider == "openai":
+        from hanzo_insights.ai.openai.openai_converter import format_openai_response
+
+        return format_openai_response(response)
+    elif provider == "gemini":
+        from hanzo_insights.ai.gemini.gemini_converter import format_gemini_response
+
+        return format_gemini_response(response)
+    return []
+
+
+def extract_available_tool_calls(provider: str, kwargs: Dict[str, Any]):
+    """
+    Extract available tool calls for the given provider.
+    """
+    if provider == "anthropic":
+        from hanzo_insights.ai.anthropic.anthropic_converter import extract_anthropic_tools
+
+        return extract_anthropic_tools(kwargs)
+    elif provider == "gemini":
+        from hanzo_insights.ai.gemini.gemini_converter import extract_gemini_tools
+
+        return extract_gemini_tools(kwargs)
+    elif provider == "openai":
+        from hanzo_insights.ai.openai.openai_converter import extract_openai_tools
+
+        return extract_openai_tools(kwargs)
+    return None
+
+
+def merge_system_prompt(
+    kwargs: Dict[str, Any], provider: str
+) -> List[FormattedMessage]:
+    """
+    Merge system prompts and format messages for the given provider.
+    """
+    if provider == "anthropic":
+        from hanzo_insights.ai.anthropic.anthropic_converter import format_anthropic_input
+
+        messages = kwargs.get("messages") or []
+        system = kwargs.get("system")
+        return format_anthropic_input(messages, system)
+    elif provider == "gemini":
+        from hanzo_insights.ai.gemini.gemini_converter import format_gemini_input_with_system
+
+        contents = kwargs.get("contents", [])
+        config = kwargs.get("config")
+        return format_gemini_input_with_system(contents, config)
+    elif provider == "openai":
+        from hanzo_insights.ai.openai.openai_converter import format_openai_input
+
+        # For OpenAI, handle both Chat Completions and Responses API
+        messages_param = kwargs.get("messages")
+        input_param = kwargs.get("input")
+
+        # Get base formatted messages
+        messages = format_openai_input(messages_param, input_param)
+
+        # Check if system prompt is provided as a separate parameter
+        if kwargs.get("system") is not None:
+            has_system = any(msg.get("role") == "system" for msg in messages)
+            if not has_system:
+                system_msg = cast(
+                    FormattedMessage,
+                    {"role": "system", "content": kwargs.get("system")},
+                )
+                messages = [system_msg] + messages
+
+        # For Responses API, add instructions to the system prompt if provided
+        if kwargs.get("instructions") is not None:
+            # Find the system message if it exists
+            system_idx = next(
+                (i for i, msg in enumerate(messages) if msg.get("role") == "system"),
+                None,
+            )
+
+            if system_idx is not None:
+                # Append instructions to existing system message
+                system_content = messages[system_idx].get("content", "")
+                messages[system_idx]["content"] = (
+                    f"{system_content}\n\n{kwargs.get('instructions')}"
+                )
+            else:
+                # Create a new system message with instructions
+                instruction_msg = cast(
+                    FormattedMessage,
+                    {"role": "system", "content": kwargs.get("instructions")},
+                )
+                messages = [instruction_msg] + messages
+
+        return messages
+
+    # Default case - return empty list
+    return []
+
+
+def call_llm_and_track_usage(
+    insights_distinct_id: Optional[str],
+    ph_client: InsightsClient,
+    provider: str,
+    insights_trace_id: Optional[str],
+    insights_properties: Optional[Dict[str, Any]],
+    insights_privacy_mode: bool,
+    insights_groups: Optional[Dict[str, Any]],
+    base_url: str,
+    call_method: Callable[..., Any],
+    **kwargs: Any,
+) -> Any:
+    """
+    Common usage-tracking logic for both sync and async calls.
+    call_method: the llm call method (e.g. openai.chat.completions.create)
+    """
+    start_time = time.time()
+    response = None
+    error = None
+    http_status = 200
+    usage: TokenUsage = TokenUsage()
+    error_params: Dict[str, Any] = {}
+
+    with new_context(client=ph_client, capture_exceptions=False):
+        if insights_distinct_id:
+            identify_context(insights_distinct_id)
+
+        try:
+            response = call_method(**kwargs)
+        except Exception as exc:
+            error = exc
+            http_status = getattr(
+                exc, "status_code", 0
+            )  # default to 0 becuase its likely an SDK error
+            error_params = {
+                "$ai_is_error": True,
+                "$ai_error": exc.__str__(),
+            }
+            # TODO: Add exception capture for OpenAI/Anthropic/Gemini wrappers when
+            # enable_exception_autocapture is True, similar to LangChain callbacks.
+            # See _capture_exception_and_update_properties in langchain/callbacks.py
+        finally:
+            end_time = time.time()
+            latency = end_time - start_time
+
+            if insights_trace_id is None:
+                insights_trace_id = str(uuid.uuid4())
+
+            # Check if we have a real user distinct_id (from param or outer context)
+            has_person_distinct_id = (
+                insights_distinct_id is not None
+                or contexts.get_context_distinct_id() is not None
+            )
+
+            if not has_person_distinct_id:
+                # Fall back to trace_id as distinct_id when no real user id is available.
+                identify_context(insights_trace_id)
+
+            if response and (
+                hasattr(response, "usage")
+                or (provider == "gemini" and hasattr(response, "usage_metadata"))
+            ):
+                usage = get_usage(response, provider)
+
+            messages = merge_system_prompt(kwargs, provider)
+            sanitized_messages = sanitize_messages(messages, provider)
+
+            tag("$ai_provider", provider)
+            tag("$ai_model", kwargs.get("model") or getattr(response, "model", None))
+            tag("$ai_model_parameters", get_model_params(kwargs))
+            tag(
+                "$ai_input",
+                with_privacy_mode(ph_client, insights_privacy_mode, sanitized_messages),
+            )
+            tag(
+                "$ai_output_choices",
+                with_privacy_mode(
+                    ph_client, insights_privacy_mode, format_response(response, provider)
+                ),
+            )
+            tag("$ai_http_status", http_status)
+            tag("$ai_input_tokens", usage.get("input_tokens", 0))
+            tag("$ai_output_tokens", usage.get("output_tokens", 0))
+            tag("$ai_latency", latency)
+            tag("$ai_trace_id", insights_trace_id)
+            tag("$ai_base_url", str(base_url))
+
+            available_tool_calls = extract_available_tool_calls(provider, kwargs)
+
+            if available_tool_calls:
+                tag("$ai_tools", available_tool_calls)
+
+            cache_read = usage.get("cache_read_input_tokens")
+            if cache_read is not None and cache_read > 0:
+                tag("$ai_cache_read_input_tokens", cache_read)
+
+            cache_creation = usage.get("cache_creation_input_tokens")
+            if cache_creation is not None and cache_creation > 0:
+                tag("$ai_cache_creation_input_tokens", cache_creation)
+
+            reasoning = usage.get("reasoning_tokens")
+            if reasoning is not None and reasoning > 0:
+                tag("$ai_reasoning_tokens", reasoning)
+
+            web_search_count = usage.get("web_search_count")
+            if web_search_count is not None and web_search_count > 0:
+                tag("$ai_web_search_count", web_search_count)
+
+            raw_usage = usage.get("raw_usage")
+            if raw_usage is not None:
+                # Already serialized by converters
+                tag("$ai_usage", raw_usage)
+
+            if not has_person_distinct_id:
+                tag("$process_person_profile", False)
+
+            # Process instructions for Responses API
+            if provider == "openai" and kwargs.get("instructions") is not None:
+                tag(
+                    "$ai_instructions",
+                    with_privacy_mode(
+                        ph_client, insights_privacy_mode, kwargs.get("instructions")
+                    ),
+                )
+
+            # send the event to Insights
+            if hasattr(ph_client, "capture") and callable(ph_client.capture):
+                sdk_tags = get_tags()
+                merged_properties = {
+                    **sdk_tags,
+                    **(insights_properties or {}),
+                    **(error_params or {}),
+                }
+                merged_properties["$ai_tokens_source"] = _get_tokens_source(
+                    sdk_tags, insights_properties
+                )
+                ph_client.capture(
+                    distinct_id=contexts.get_context_distinct_id(),
+                    event="$ai_generation",
+                    properties=merged_properties,
+                    groups=insights_groups,
+                )
+
+        if error:
+            raise error
+
+    return response
+
+
+async def call_llm_and_track_usage_async(
+    insights_distinct_id: Optional[str],
+    ph_client: InsightsClient,
+    provider: str,
+    insights_trace_id: Optional[str],
+    insights_properties: Optional[Dict[str, Any]],
+    insights_privacy_mode: bool,
+    insights_groups: Optional[Dict[str, Any]],
+    base_url: str,
+    call_async_method: Callable[..., Any],
+    **kwargs: Any,
+) -> Any:
+    start_time = time.time()
+    response = None
+    error = None
+    http_status = 200
+    usage: TokenUsage = TokenUsage()
+    error_params: Dict[str, Any] = {}
+
+    with new_context(client=ph_client, capture_exceptions=False):
+        if insights_distinct_id:
+            identify_context(insights_distinct_id)
+
+        try:
+            response = await call_async_method(**kwargs)
+        except Exception as exc:
+            error = exc
+            http_status = getattr(
+                exc, "status_code", 0
+            )  # default to 0 because its likely an SDK error
+            error_params = {
+                "$ai_is_error": True,
+                "$ai_error": exc.__str__(),
+            }
+            # TODO: Add exception capture for OpenAI/Anthropic/Gemini wrappers when
+            # enable_exception_autocapture is True, similar to LangChain callbacks.
+            # See _capture_exception_and_update_properties in langchain/callbacks.py
+        finally:
+            end_time = time.time()
+            latency = end_time - start_time
+
+            if insights_trace_id is None:
+                insights_trace_id = str(uuid.uuid4())
+
+            # Check if we have a real user distinct_id (from param or outer context)
+            has_person_distinct_id = (
+                insights_distinct_id is not None
+                or contexts.get_context_distinct_id() is not None
+            )
+
+            if not has_person_distinct_id:
+                # Fall back to trace_id as distinct_id when no real user id is available.
+                identify_context(insights_trace_id)
+
+            if response and (
+                hasattr(response, "usage")
+                or (provider == "gemini" and hasattr(response, "usage_metadata"))
+            ):
+                usage = get_usage(response, provider)
+
+            messages = merge_system_prompt(kwargs, provider)
+            sanitized_messages = sanitize_messages(messages, provider)
+
+            tag("$ai_provider", provider)
+            tag("$ai_model", kwargs.get("model") or getattr(response, "model", None))
+            tag("$ai_model_parameters", get_model_params(kwargs))
+            tag(
+                "$ai_input",
+                with_privacy_mode(ph_client, insights_privacy_mode, sanitized_messages),
+            )
+            tag(
+                "$ai_output_choices",
+                with_privacy_mode(
+                    ph_client, insights_privacy_mode, format_response(response, provider)
+                ),
+            )
+            tag("$ai_http_status", http_status)
+            tag("$ai_input_tokens", usage.get("input_tokens", 0))
+            tag("$ai_output_tokens", usage.get("output_tokens", 0))
+            tag("$ai_latency", latency)
+            tag("$ai_trace_id", insights_trace_id)
+            tag("$ai_base_url", str(base_url))
+
+            available_tool_calls = extract_available_tool_calls(provider, kwargs)
+
+            if available_tool_calls:
+                tag("$ai_tools", available_tool_calls)
+
+            cache_read = usage.get("cache_read_input_tokens")
+            if cache_read is not None and cache_read > 0:
+                tag("$ai_cache_read_input_tokens", cache_read)
+
+            cache_creation = usage.get("cache_creation_input_tokens")
+            if cache_creation is not None and cache_creation > 0:
+                tag("$ai_cache_creation_input_tokens", cache_creation)
+
+            reasoning = usage.get("reasoning_tokens")
+            if reasoning is not None and reasoning > 0:
+                tag("$ai_reasoning_tokens", reasoning)
+
+            web_search_count = usage.get("web_search_count")
+            if web_search_count is not None and web_search_count > 0:
+                tag("$ai_web_search_count", web_search_count)
+
+            raw_usage = usage.get("raw_usage")
+            if raw_usage is not None:
+                # Already serialized by converters
+                tag("$ai_usage", raw_usage)
+
+            if not has_person_distinct_id:
+                tag("$process_person_profile", False)
+
+            # Process instructions for Responses API
+            if provider == "openai" and kwargs.get("instructions") is not None:
+                tag(
+                    "$ai_instructions",
+                    with_privacy_mode(
+                        ph_client, insights_privacy_mode, kwargs.get("instructions")
+                    ),
+                )
+
+            # send the event to Insights
+            if hasattr(ph_client, "capture") and callable(ph_client.capture):
+                sdk_tags = get_tags()
+                merged_properties = {
+                    **sdk_tags,
+                    **(insights_properties or {}),
+                    **(error_params or {}),
+                }
+                merged_properties["$ai_tokens_source"] = _get_tokens_source(
+                    sdk_tags, insights_properties
+                )
+                ph_client.capture(
+                    distinct_id=contexts.get_context_distinct_id(),
+                    event="$ai_generation",
+                    properties=merged_properties,
+                    groups=insights_groups,
+                )
+
+        if error:
+            raise error
+
+    return response
+
+
+def sanitize_messages(data: Any, provider: str) -> Any:
+    """Sanitize messages using provider-specific sanitization functions."""
+    if provider == "anthropic":
+        return sanitize_anthropic(data)
+    elif provider == "openai":
+        return sanitize_openai(data)
+    elif provider == "gemini":
+        return sanitize_gemini(data)
+    elif provider == "langchain":
+        return sanitize_langchain(data)
+    return data
+
+
+def with_privacy_mode(ph_client: InsightsClient, privacy_mode: bool, value: Any):
+    if ph_client.privacy_mode or privacy_mode:
+        return None
+    return value
+
+
+def capture_streaming_event(
+    ph_client: InsightsClient,
+    event_data: StreamingEventData,
+):
+    """
+    Unified streaming event capture for all LLM providers.
+
+    This function handles the common logic for capturing streaming events across all providers.
+    All provider-specific formatting should be done BEFORE calling this function.
+
+    The function handles:
+    - Building Insights event properties
+    - Extracting and adding tools based on provider
+    - Applying privacy mode
+    - Adding special token fields (cache, reasoning)
+    - Provider-specific fields (e.g., OpenAI instructions)
+    - Sending the event to Insights
+
+    Args:
+        ph_client: Insights client instance
+        event_data: Standardized streaming event data containing all necessary information
+    """
+    trace_id = event_data.get("trace_id") or str(uuid.uuid4())
+
+    # Build base event properties
+    event_properties = {
+        "$ai_provider": event_data["provider"],
+        "$ai_model": event_data["model"],
+        "$ai_model_parameters": get_model_params(event_data["kwargs"]),
+        "$ai_input": with_privacy_mode(
+            ph_client,
+            event_data["privacy_mode"],
+            event_data["formatted_input"],
+        ),
+        "$ai_output_choices": with_privacy_mode(
+            ph_client,
+            event_data["privacy_mode"],
+            event_data["formatted_output"],
+        ),
+        "$ai_http_status": 200,
+        "$ai_input_tokens": event_data["usage_stats"].get("input_tokens", 0),
+        "$ai_output_tokens": event_data["usage_stats"].get("output_tokens", 0),
+        "$ai_latency": event_data["latency"],
+        "$ai_trace_id": trace_id,
+        "$ai_base_url": str(event_data["base_url"]),
+        **(event_data.get("properties") or {}),
+    }
+
+    # Determine token source: SDK-computed vs externally overridden
+    sdk_token_tags = {
+        "$ai_input_tokens": event_data["usage_stats"].get("input_tokens", 0),
+        "$ai_output_tokens": event_data["usage_stats"].get("output_tokens", 0),
+    }
+    event_properties["$ai_tokens_source"] = _get_tokens_source(
+        sdk_token_tags, event_data.get("properties")
+    )
+
+    # Extract and add tools based on provider
+    available_tools = extract_available_tool_calls(
+        event_data["provider"],
+        event_data["kwargs"],
+    )
+    if available_tools:
+        event_properties["$ai_tools"] = available_tools
+
+    # Add optional token fields
+    # For Anthropic, always include cache fields even if 0 (backward compatibility)
+    # For others, only include if present and non-zero
+    if event_data["provider"] == "anthropic":
+        # Anthropic always includes cache fields
+        cache_read = event_data["usage_stats"].get("cache_read_input_tokens", 0)
+        cache_creation = event_data["usage_stats"].get("cache_creation_input_tokens", 0)
+        event_properties["$ai_cache_read_input_tokens"] = cache_read
+        event_properties["$ai_cache_creation_input_tokens"] = cache_creation
+    else:
+        # Other providers only include if non-zero
+        optional_token_fields = [
+            "cache_read_input_tokens",
+            "cache_creation_input_tokens",
+            "reasoning_tokens",
+        ]
+
+        for field in optional_token_fields:
+            value = event_data["usage_stats"].get(field)
+            if value is not None and isinstance(value, int) and value > 0:
+                event_properties[f"$ai_{field}"] = value
+
+    # Add web search count if present (all providers)
+    web_search_count = event_data["usage_stats"].get("web_search_count")
+    if (
+        web_search_count is not None
+        and isinstance(web_search_count, int)
+        and web_search_count > 0
+    ):
+        event_properties["$ai_web_search_count"] = web_search_count
+
+    # Add raw usage metadata if present (all providers)
+    raw_usage = event_data["usage_stats"].get("raw_usage")
+    if raw_usage is not None:
+        # Already serialized by converters
+        event_properties["$ai_usage"] = raw_usage
+
+    # Handle provider-specific fields
+    if (
+        event_data["provider"] == "openai"
+        and event_data["kwargs"].get("instructions") is not None
+    ):
+        event_properties["$ai_instructions"] = with_privacy_mode(
+            ph_client,
+            event_data["privacy_mode"],
+            event_data["kwargs"]["instructions"],
+        )
+
+    if event_data.get("distinct_id") is None:
+        event_properties["$process_person_profile"] = False
+
+    # Send event to Insights
+    if hasattr(ph_client, "capture"):
+        ph_client.capture(
+            distinct_id=event_data.get("distinct_id") or trace_id,
+            event="$ai_generation",
+            properties=event_properties,
+            groups=event_data.get("groups"),
+        )

hanzo_insights/args.py

@@ -5,7 +5,7 @@ from datetime import datetime
 import numbers
 from uuid import UUID
 
-from posthog.types import SendFeatureFlagsOptions
+from hanzo_insights.types import SendFeatureFlagsOptions
 
 ID_TYPES = Union[numbers.Number, str, UUID, int]
 

hanzo_insights/client.py

@@ -11,19 +11,20 @@ from dateutil.tz import tzutc
 from six import string_types
 from typing_extensions import Unpack
 
-from posthog.args import ID_TYPES, ExceptionArg, OptionalCaptureArgs, OptionalSetArgs
-from posthog.consumer import Consumer
-from posthog.contexts import (
+from hanzo_insights.args import ID_TYPES, ExceptionArg, OptionalCaptureArgs, OptionalSetArgs
+from hanzo_insights.consumer import Consumer
+from hanzo_insights.contexts import (
     _get_current_context,
     get_capture_exception_code_variables_context,
     get_code_variables_ignore_patterns_context,
     get_code_variables_mask_patterns_context,
+    get_context_device_id,
     get_context_distinct_id,
     get_context_session_id,
     new_context,
 )
-from posthog.exception_capture import ExceptionCapture
-from posthog.exception_utils import (
+from hanzo_insights.exception_capture import ExceptionCapture
+from hanzo_insights.exception_utils import (
     DEFAULT_CODE_VARIABLES_IGNORE_PATTERNS,
     DEFAULT_CODE_VARIABLES_MASK_PATTERNS,
     exc_info_from_error,
@@ -33,17 +34,18 @@ from posthog.exception_utils import (
     mark_exception_as_captured,
     try_attach_code_variables_to_frames,
 )
-from posthog.feature_flags import (
+from hanzo_insights.feature_flags import (
     InconclusiveMatchError,
     RequiresServerEvaluation,
     match_feature_flag_properties,
+    resolve_bucketing_value,
 )
-from posthog.flag_definition_cache import (
+from hanzo_insights.flag_definition_cache import (
     FlagDefinitionCacheData,
     FlagDefinitionCacheProvider,
 )
-from posthog.poller import Poller
-from posthog.request import (
+from hanzo_insights.poller import Poller
+from hanzo_insights.request import (
     DEFAULT_HOST,
     APIError,
     QuotaLimitError,
@@ -55,7 +57,7 @@ from posthog.request import (
     get,
     remote_config,
 )
-from posthog.types import (
+from hanzo_insights.types import (
     FeatureFlag,
     FeatureFlagError,
     FeatureFlagResult,
@@ -69,7 +71,7 @@ from posthog.types import (
     to_payloads,
     to_values,
 )
-from posthog.utils import (
+from hanzo_insights.utils import (
     FlagCache,
     RedisFlagCache,
     SizeLimitedDict,
@@ -77,7 +79,7 @@ from posthog.utils import (
     guess_timezone,
     system_context,
 )
-from posthog.version import VERSION
+from hanzo_insights.version import VERSION
 
 try:
     import queue
@@ -147,22 +149,22 @@ def no_throw(default_return=None):
 
 class Client(object):
     """
-    This is the SDK reference for the PostHog Python SDK.
+    This is the SDK reference for the Hanzo Insights Python SDK.
     You can learn more about example usage in the [Python SDK documentation](/docs/libraries/python).
     You can also follow [Flask](/docs/libraries/flask) and [Django](/docs/libraries/django)
-    guides to integrate PostHog into your project.
+    guides to integrate Insights into your project.
 
     Examples:
         ```python
-        from posthog import Posthog
-        posthog = Posthog('<ph_project_api_key>', host='<ph_client_api_host>')
-        posthog.debug = True
+        from hanzo_insights import Insights
+        client = Insights('<ph_project_api_key>', host='<ph_client_api_host>')
+        hanzo_insights.debug = True
         if settings.TEST:
-            posthog.disabled = True
+            hanzo_insights.disabled = True
         ```
     """
 
-    log = logging.getLogger("posthog")
+    log = logging.getLogger("hanzo_insights")
 
     def __init__(
         self,
@@ -200,7 +202,7 @@ class Client(object):
         in_app_modules: list[str] | None = None,
     ):
         """
-        Initialize a new PostHog client instance.
+        Initialize a new Insights client instance.
 
         Args:
             project_api_key: The project API key.
@@ -209,9 +211,9 @@ class Client(object):
 
         Examples:
             ```python
-            from posthog import Posthog
+            from hanzo_insights import Insights
 
-            posthog = Posthog('<ph_project_api_key>', host='<ph_app_host>')
+            client = Insights('<ph_project_api_key>', host='<ph_app_host>')
             ```
 
         Category:
@@ -340,9 +342,9 @@ class Client(object):
 
         Examples:
             ```python
-            with posthog.new_context():
+            with hanzo_insights.new_context():
                 identify_context('<distinct_id>')
-                posthog.capture('event_name')
+                hanzo_insights.capture('event_name')
             ```
 
         Category:
@@ -382,6 +384,7 @@ class Client(object):
         group_properties=None,
         disable_geoip=None,
         flag_keys_to_evaluate: Optional[list[str]] = None,
+        device_id: Optional[str] = None,
     ) -> dict[str, Union[bool, str]]:
         """
         Get feature flag variants for a user by calling decide.
@@ -394,6 +397,7 @@ class Client(object):
             disable_geoip: Whether to disable GeoIP for this request.
             flag_keys_to_evaluate: A list of specific flag keys to evaluate. If provided,
                 only these flags will be evaluated, improving performance.
+            device_id: The device ID for this request.
 
         Category:
             Feature flags
@@ -405,6 +409,7 @@ class Client(object):
             group_properties,
             disable_geoip,
             flag_keys_to_evaluate,
+            device_id=device_id,
         )
         return to_values(resp_data) or {}
 
@@ -416,6 +421,7 @@ class Client(object):
         group_properties=None,
         disable_geoip=None,
         flag_keys_to_evaluate: Optional[list[str]] = None,
+        device_id: Optional[str] = None,
     ) -> dict[str, str]:
         """
         Get feature flag payloads for a user by calling decide.
@@ -428,10 +434,11 @@ class Client(object):
             disable_geoip: Whether to disable GeoIP for this request.
             flag_keys_to_evaluate: A list of specific flag keys to evaluate. If provided,
                 only these flags will be evaluated, improving performance.
+            device_id: The device ID for this request.
 
         Examples:
             ```python
-            payloads = posthog.get_feature_payloads('<distinct_id>')
+            payloads = hanzo_insights.get_feature_payloads('<distinct_id>')
             ```
 
         Category:
@@ -444,6 +451,7 @@ class Client(object):
             group_properties,
             disable_geoip,
             flag_keys_to_evaluate,
+            device_id=device_id,
         )
         return to_payloads(resp_data) or {}
 
@@ -455,6 +463,7 @@ class Client(object):
         group_properties=None,
         disable_geoip=None,
         flag_keys_to_evaluate: Optional[list[str]] = None,
+        device_id: Optional[str] = None,
     ) -> FlagsAndPayloads:
         """
         Get feature flags and payloads for a user by calling decide.
@@ -467,10 +476,11 @@ class Client(object):
             disable_geoip: Whether to disable GeoIP for this request.
             flag_keys_to_evaluate: A list of specific flag keys to evaluate. If provided,
                 only these flags will be evaluated, improving performance.
+            device_id: The device ID for this request.
 
         Examples:
             ```python
-            result = posthog.get_feature_flags_and_payloads('<distinct_id>')
+            result = hanzo_insights.get_feature_flags_and_payloads('<distinct_id>')
             ```
 
         Category:
@@ -483,6 +493,7 @@ class Client(object):
             group_properties,
             disable_geoip,
             flag_keys_to_evaluate,
+            device_id=device_id,
         )
         return to_flags_and_payloads(resp)
 
@@ -494,6 +505,7 @@ class Client(object):
         group_properties=None,
         disable_geoip=None,
         flag_keys_to_evaluate: Optional[list[str]] = None,
+        device_id: Optional[str] = None,
     ) -> FlagsResponse:
         """
         Get feature flags decision.
@@ -506,10 +518,11 @@ class Client(object):
             disable_geoip: Whether to disable GeoIP for this request.
             flag_keys_to_evaluate: A list of specific flag keys to evaluate. If provided,
                 only these flags will be evaluated, improving performance.
+            device_id: The device ID for this request.
 
         Examples:
             ```python
-            decision = posthog.get_flags_decision('user123')
+            decision = hanzo_insights.get_flags_decision('user123')
             ```
 
         Category:
@@ -522,6 +535,9 @@ class Client(object):
         if distinct_id is None:
             distinct_id = get_context_distinct_id()
 
+        if device_id is None:
+            device_id = get_context_device_id()
+
         if disable_geoip is None:
             disable_geoip = self.disable_geoip
 
@@ -534,6 +550,7 @@ class Client(object):
             "person_properties": person_properties,
             "group_properties": group_properties,
             "geoip_disable": disable_geoip,
+            "device_id": device_id,
         }
 
         if flag_keys_to_evaluate:
@@ -553,7 +570,7 @@ class Client(object):
         self, event: str, **kwargs: Unpack[OptionalCaptureArgs]
     ) -> Optional[str]:
         """
-        Captures an event manually. [Learn about capture best practices](https://posthog.com/docs/product-analytics/capture-events)
+        Captures an event manually. [Learn about capture best practices](https://insights.hanzo.ai/docs/product-analytics/capture-events)
 
         Args:
             event: The event name to capture.
@@ -568,20 +585,20 @@ class Client(object):
         Examples:
             ```python
             # Anonymous event
-            posthog.capture('some-anon-event')
+            hanzo_insights.capture('some-anon-event')
             ```
             ```python
             # Context usage
-            from posthog import identify_context, new_context
+            from hanzo_insights import identify_context, new_context
             with new_context():
                 identify_context('distinct_id_of_the_user')
-                posthog.capture('user_signed_up')
-                posthog.capture('user_logged_in')
-                posthog.capture('some-custom-action', distinct_id='distinct_id_of_the_user')
+                hanzo_insights.capture('user_signed_up')
+                hanzo_insights.capture('user_logged_in')
+                hanzo_insights.capture('some-custom-action', distinct_id='distinct_id_of_the_user')
             ```
             ```python
             # Set event properties
-            posthog.capture(
+            hanzo_insights.capture(
                 "user_signed_up",
                 distinct_id="distinct_id_of_the_user",
                 properties={
@@ -592,7 +609,7 @@ class Client(object):
             ```
             ```python
             # Page view event
-            posthog.capture('$pageview', distinct_id="distinct_id_of_the_user", properties={'$current_url': 'https://example.com'})
+            hanzo_insights.capture('$pageview', distinct_id="distinct_id_of_the_user", properties={'$current_url': 'https://example.com'})
             ```
 
         Category:
@@ -752,7 +769,7 @@ class Client(object):
         Examples:
             ```python
             # Set with distinct id
-            posthog.set(distinct_id='user123', properties={'name': 'Max Hedgehog'})
+            hanzo_insights.set(distinct_id='user123', properties={'name': 'Max Hedgehog'})
             ```
 
         Category:
@@ -799,7 +816,7 @@ class Client(object):
 
         Examples:
             ```python
-            posthog.set_once(distinct_id='user123', properties={'initial_signup_date': '2024-01-01'})
+            hanzo_insights.set_once(distinct_id='user123', properties={'initial_signup_date': '2024-01-01'})
             ```
 
         Category:
@@ -856,7 +873,7 @@ class Client(object):
 
         Examples:
             ```python
-            posthog.group_identify('company', 'company_id_in_your_db', {
+            hanzo_insights.group_identify('company', 'company_id_in_your_db', {
                 'name': 'Awesome Inc.',
                 'employees': 11
             })
@@ -911,7 +928,7 @@ class Client(object):
 
         Examples:
             ```python
-            posthog.alias(previous_id='distinct_id', distinct_id='alias_id')
+            hanzo_insights.alias(previous_id='distinct_id', distinct_id='alias_id')
             ```
 
         Category:
@@ -961,7 +978,7 @@ class Client(object):
                 # Some code that might fail
                 pass
             except Exception as e:
-                posthog.capture_exception(e, 'user_distinct_id', properties=additional_properties)
+                hanzo_insights.capture_exception(e, 'user_distinct_id', properties=additional_properties)
             ```
 
         Category:
@@ -1091,7 +1108,7 @@ class Client(object):
 
         if not msg.get("properties"):
             msg["properties"] = {}
-        msg["properties"]["$lib"] = "posthog-python"
+        msg["properties"]["$lib"] = "insights-python"
         msg["properties"]["$lib_version"] = VERSION
 
         if disable_geoip is None:
@@ -1151,8 +1168,8 @@ class Client(object):
 
         Examples:
             ```python
-            posthog.capture('event_name')
-            posthog.flush()  # Ensures the event is sent immediately
+            hanzo_insights.capture('event_name')
+            hanzo_insights.flush()  # Ensures the event is sent immediately
             ```
         """
         queue = self.queue
@@ -1167,7 +1184,7 @@ class Client(object):
 
         Examples:
             ```python
-            posthog.join()
+            hanzo_insights.join()
             ```
         """
         if self.consumers:
@@ -1195,7 +1212,7 @@ class Client(object):
 
         Examples:
             ```python
-            posthog.shutdown()
+            hanzo_insights.shutdown()
             ```
         """
         self.flush()
@@ -1268,7 +1285,7 @@ class Client(object):
             self._fetch_feature_flags_from_api()
 
     def _fetch_feature_flags_from_api(self):
-        """Fetch feature flags from the PostHog API."""
+        """Fetch feature flags from the Insights API."""
         try:
             # Store old flags to detect changes
             old_flags_by_key: dict[str, dict] = self.feature_flags_by_key or {}
@@ -1317,18 +1334,25 @@ class Client(object):
         except APIError as e:
             if e.status == 401:
                 self.log.error(
-                    "[FEATURE FLAGS] Error loading feature flags: To use feature flags, please set a valid personal_api_key. More information: https://posthog.com/docs/api/overview"
+                    "[FEATURE FLAGS] Error loading feature flags: To use feature flags, please set a valid personal_api_key. More information: https://insights.hanzo.ai/docs/api/overview"
                 )
+                self.feature_flags = []
+                self.group_type_mapping = {}
+                self.cohorts = {}
+
+                if self.flag_cache:
+                    self.flag_cache.clear()
+
                 if self.debug:
                     raise APIError(
                         status=401,
                         message="You are using a write-only key with feature flags. "
                         "To use feature flags, please set a personal_api_key "
-                        "More information: https://posthog.com/docs/api/overview",
+                        "More information: https://insights.hanzo.ai/docs/api/overview",
                     )
             elif e.status == 402:
                 self.log.warning(
-                    "[FEATURE FLAGS] PostHog feature flags quota limited, resetting feature flag data.  Learn more about billing limits at https://posthog.com/docs/billing/limits-alerts"
+                    "[FEATURE FLAGS] Insights feature flags quota limited, resetting feature flag data.  Learn more about billing limits at https://insights.hanzo.ai/docs/billing/limits-alerts"
                 )
                 # Reset all feature flag data when quota limited
                 self.feature_flags = []
@@ -1342,7 +1366,7 @@ class Client(object):
                 if self.debug:
                     raise APIError(
                         status=402,
-                        message="PostHog feature flags quota limited",
+                        message="Insights feature flags quota limited",
                     )
             else:
                 self.log.error(f"[FEATURE FLAGS] Error loading feature flags: {e}")
@@ -1361,7 +1385,7 @@ class Client(object):
 
         Examples:
             ```python
-            posthog.load_feature_flags()
+            hanzo_insights.load_feature_flags()
             ```
 
         Category:
@@ -1395,6 +1419,7 @@ class Client(object):
         person_properties=None,
         group_properties=None,
         warn_on_unknown_groups=True,
+        device_id=None,
     ) -> FlagValue:
         groups = groups or {}
         person_properties = person_properties or {}
@@ -1435,22 +1460,35 @@ class Client(object):
                     )
                 return False
 
+            if group_name not in group_properties:
+                raise InconclusiveMatchError(
+                    f"Flag has no group properties for group '{group_name}'"
+                )
             focused_group_properties = group_properties[group_name]
+            group_key = groups[group_name]
             return match_feature_flag_properties(
                 feature_flag,
-                groups[group_name],
+                group_key,
                 focused_group_properties,
-                self.feature_flags_by_key,
-                evaluation_cache,
+                cohort_properties=self.cohorts,
+                flags_by_key=self.feature_flags_by_key,
+                evaluation_cache=evaluation_cache,
+                device_id=device_id,
+                bucketing_value=group_key,
             )
         else:
+            bucketing_value = resolve_bucketing_value(
+                feature_flag, distinct_id, device_id
+            )
             return match_feature_flag_properties(
                 feature_flag,
                 distinct_id,
                 person_properties,
-                self.cohorts,
-                self.feature_flags_by_key,
-                evaluation_cache,
+                cohort_properties=self.cohorts,
+                flags_by_key=self.feature_flags_by_key,
+                evaluation_cache=evaluation_cache,
+                device_id=device_id,
+                bucketing_value=bucketing_value,
             )
 
     def feature_enabled(
@@ -1464,6 +1502,7 @@ class Client(object):
         only_evaluate_locally=False,
         send_feature_flag_events=True,
         disable_geoip=None,
+        device_id: Optional[str] = None,
     ):
         """
         Check if a feature flag is enabled for a user.
@@ -1477,14 +1516,15 @@ class Client(object):
             only_evaluate_locally: Whether to only evaluate locally.
             send_feature_flag_events: Whether to send feature flag events.
             disable_geoip: Whether to disable GeoIP for this request.
+            device_id: The device ID for this request.
 
         Examples:
             ```python
-            is_my_flag_enabled = posthog.feature_enabled('flag-key', 'distinct_id_of_your_user')
+            is_my_flag_enabled = hanzo_insights.feature_enabled('flag-key', 'distinct_id_of_your_user')
             if is_my_flag_enabled:
                 # Do something differently for this user
                 # Optional: fetch the payload
-                matched_flag_payload = posthog.get_feature_flag_payload('flag-key', 'distinct_id_of_your_user')
+                matched_flag_payload = hanzo_insights.get_feature_flag_payload('flag-key', 'distinct_id_of_your_user')
             ```
 
         Category:
@@ -1499,6 +1539,7 @@ class Client(object):
             only_evaluate_locally=only_evaluate_locally,
             send_feature_flag_events=send_feature_flag_events,
             disable_geoip=disable_geoip,
+            device_id=device_id,
         )
 
         if response is None:
@@ -1530,6 +1571,7 @@ class Client(object):
         only_evaluate_locally=False,
         send_feature_flag_events=True,
         disable_geoip=None,
+        device_id: Optional[str] = None,
     ) -> Optional[FeatureFlagResult]:
         if self.disabled:
             return None
@@ -1553,8 +1595,12 @@ class Client(object):
         evaluated_at = None
         feature_flag_error: Optional[str] = None
 
+        # Resolve device_id from context if not provided
+        if device_id is None:
+            device_id = get_context_device_id()
+
         flag_value = self._locally_evaluate_flag(
-            key, distinct_id, groups, person_properties, group_properties
+            key, distinct_id, groups, person_properties, group_properties, device_id
         )
         flag_was_locally_evaluated = flag_value is not None
 
@@ -1574,7 +1620,13 @@ class Client(object):
                 self.flag_cache.set_cached_flag(
                     distinct_id, key, flag_result, self.flag_definition_version
                 )
-        elif not only_evaluate_locally:
+        elif only_evaluate_locally:
+            if self.feature_flags is None:
+                self.log.warning(
+                    "[FEATURE FLAGS] Local evaluation called but feature flag definitions are not loaded yet. "
+                    "Returning None. You can call load_feature_flags() to load flags explicitly."
+                )
+        else:
             try:
                 flag_details, request_id, evaluated_at, errors_while_computing = (
                     self._get_feature_flag_details_from_server(
@@ -1584,6 +1636,7 @@ class Client(object):
                         person_properties,
                         group_properties,
                         disable_geoip,
+                        device_id=device_id,
                     )
                 )
                 errors = []
@@ -1656,6 +1709,7 @@ class Client(object):
         only_evaluate_locally=False,
         send_feature_flag_events=True,
         disable_geoip=None,
+        device_id: Optional[str] = None,
     ) -> Optional[FeatureFlagResult]:
         """
         Get a FeatureFlagResult object which contains the flag result and payload for a key by evaluating locally or remotely
@@ -1664,7 +1718,7 @@ class Client(object):
 
         Examples:
             ```python
-            flag_result = posthog.get_feature_flag_result('flag-key', 'distinct_id_of_your_user')
+            flag_result = hanzo_insights.get_feature_flag_result('flag-key', 'distinct_id_of_your_user')
             if flag_result and flag_result.get_value() == 'variant-key':
                 # Do something differently for this user
                 # Optional: fetch the payload
@@ -1680,6 +1734,7 @@ class Client(object):
             only_evaluate_locally: Whether to only evaluate locally.
             send_feature_flag_events: Whether to send feature flag events.
             disable_geoip: Whether to disable GeoIP for this request.
+            device_id: The device ID for this request.
 
         Returns:
             Optional[FeatureFlagResult]: The feature flag result or None if disabled/not found.
@@ -1693,6 +1748,7 @@ class Client(object):
             only_evaluate_locally=only_evaluate_locally,
             send_feature_flag_events=send_feature_flag_events,
             disable_geoip=disable_geoip,
+            device_id=device_id,
         )
 
     def get_feature_flag(
@@ -1706,6 +1762,7 @@ class Client(object):
         only_evaluate_locally=False,
         send_feature_flag_events=True,
         disable_geoip=None,
+        device_id: Optional[str] = None,
     ) -> Optional[FlagValue]:
         """
         Get multivariate feature flag value for a user.
@@ -1719,14 +1776,15 @@ class Client(object):
             only_evaluate_locally: Whether to only evaluate locally.
             send_feature_flag_events: Whether to send feature flag events.
             disable_geoip: Whether to disable GeoIP for this request.
+            device_id: The device ID for this request.
 
         Examples:
             ```python
-            enabled_variant = posthog.get_feature_flag('flag-key', 'distinct_id_of_your_user')
+            enabled_variant = hanzo_insights.get_feature_flag('flag-key', 'distinct_id_of_your_user')
             if enabled_variant == 'variant-key': # replace 'variant-key' with the key of your variant
                 # Do something differently for this user
                 # Optional: fetch the payload
-                matched_flag_payload = posthog.get_feature_flag_payload('flag-key', 'distinct_id_of_your_user')
+                matched_flag_payload = hanzo_insights.get_feature_flag_payload('flag-key', 'distinct_id_of_your_user')
             ```
 
         Category:
@@ -1741,6 +1799,7 @@ class Client(object):
             only_evaluate_locally=only_evaluate_locally,
             send_feature_flag_events=send_feature_flag_events,
             disable_geoip=disable_geoip,
+            device_id=device_id,
         )
         return feature_flag_result.get_value() if feature_flag_result else None
 
@@ -1751,6 +1810,7 @@ class Client(object):
         groups: dict[str, str],
         person_properties: dict[str, str],
         group_properties: dict[str, str],
+        device_id: Optional[str] = None,
     ) -> Optional[FlagValue]:
         if self.feature_flags is None and self.personal_api_key:
             self.load_feature_flags()
@@ -1770,6 +1830,7 @@ class Client(object):
                         groups=groups,
                         person_properties=person_properties,
                         group_properties=group_properties,
+                        device_id=device_id,
                     )
                     self.log.debug(
                         f"Successfully computed flag locally: {key} -> {response}"
@@ -1794,6 +1855,7 @@ class Client(object):
         only_evaluate_locally=False,
         send_feature_flag_events=False,
         disable_geoip=None,
+        device_id: Optional[str] = None,
     ):
         """
         Get the payload for a feature flag.
@@ -1808,15 +1870,16 @@ class Client(object):
             only_evaluate_locally: Whether to only evaluate locally.
             send_feature_flag_events: Deprecated. Use get_feature_flag() instead if you need events.
             disable_geoip: Whether to disable GeoIP for this request.
+            device_id: The device ID for this request.
 
         Examples:
             ```python
-            is_my_flag_enabled = posthog.feature_enabled('flag-key', 'distinct_id_of_your_user')
+            is_my_flag_enabled = hanzo_insights.feature_enabled('flag-key', 'distinct_id_of_your_user')
 
             if is_my_flag_enabled:
                 # Do something differently for this user
                 # Optional: fetch the payload
-                matched_flag_payload = posthog.get_feature_flag_payload('flag-key', 'distinct_id_of_your_user')
+                matched_flag_payload = hanzo_insights.get_feature_flag_payload('flag-key', 'distinct_id_of_your_user')
             ```
 
         Category:
@@ -1840,6 +1903,7 @@ class Client(object):
             only_evaluate_locally=only_evaluate_locally,
             send_feature_flag_events=send_feature_flag_events,
             disable_geoip=disable_geoip,
+            device_id=device_id,
         )
         return feature_flag_result.payload if feature_flag_result else None
 
@@ -1851,6 +1915,7 @@ class Client(object):
         person_properties: dict[str, str],
         group_properties: dict[str, str],
         disable_geoip: Optional[bool],
+        device_id: Optional[str] = None,
     ) -> tuple[Optional[FeatureFlag], Optional[str], Optional[int], bool]:
         """
         Calls /flags and returns the flag details, request id, evaluated at timestamp,
@@ -1863,6 +1928,7 @@ class Client(object):
             group_properties,
             disable_geoip,
             flag_keys_to_evaluate=[key],
+            device_id=device_id,
         )
         request_id = resp_data.get("requestId")
         evaluated_at = resp_data.get("evaluatedAt")
@@ -1889,10 +1955,12 @@ class Client(object):
             f"{key}_{'::null::' if response is None else str(response)}"
         )
 
-        if (
-            feature_flag_reported_key
-            not in self.distinct_ids_feature_flags_reported[distinct_id]
-        ):
+        reported_flags = self.distinct_ids_feature_flags_reported.get(distinct_id)
+        if reported_flags is None:
+            reported_flags = set()
+            self.distinct_ids_feature_flags_reported[distinct_id] = reported_flags
+
+        if feature_flag_reported_key not in reported_flags:
             properties: dict[str, Any] = {
                 "$feature_flag": key,
                 "$feature_flag_response": response,
@@ -1928,9 +1996,7 @@ class Client(object):
                 groups=groups,
                 disable_geoip=disable_geoip,
             )
-            self.distinct_ids_feature_flags_reported[distinct_id].add(
-                feature_flag_reported_key
-            )
+            reported_flags.add(feature_flag_reported_key)
 
     def get_remote_config_payload(self, key: str):
         if self.disabled:
@@ -1987,6 +2053,7 @@ class Client(object):
         only_evaluate_locally=False,
         disable_geoip=None,
         flag_keys_to_evaluate: Optional[list[str]] = None,
+        device_id: Optional[str] = None,
     ) -> Optional[dict[str, Union[bool, str]]]:
         """
         Get all feature flags for a user.
@@ -2000,10 +2067,11 @@ class Client(object):
             disable_geoip: Whether to disable GeoIP for this request.
             flag_keys_to_evaluate: A list of specific flag keys to evaluate. If provided,
                 only these flags will be evaluated, improving performance.
+            device_id: The device ID for this request.
 
         Examples:
             ```python
-            posthog.get_all_flags('distinct_id_of_your_user')
+            hanzo_insights.get_all_flags('distinct_id_of_your_user')
             ```
 
         Category:
@@ -2017,6 +2085,7 @@ class Client(object):
             only_evaluate_locally=only_evaluate_locally,
             disable_geoip=disable_geoip,
             flag_keys_to_evaluate=flag_keys_to_evaluate,
+            device_id=device_id,
         )
 
         return response["featureFlags"]
@@ -2031,6 +2100,7 @@ class Client(object):
         only_evaluate_locally=False,
         disable_geoip=None,
         flag_keys_to_evaluate: Optional[list[str]] = None,
+        device_id: Optional[str] = None,
     ) -> FlagsAndPayloads:
         """
         Get all feature flags and their payloads for a user.
@@ -2044,10 +2114,11 @@ class Client(object):
             disable_geoip: Whether to disable GeoIP for this request.
             flag_keys_to_evaluate: A list of specific flag keys to evaluate. If provided,
                 only these flags will be evaluated, improving performance.
+            device_id: The device ID for this request.
 
         Examples:
             ```python
-            posthog.get_all_flags_and_payloads('distinct_id_of_your_user')
+            hanzo_insights.get_all_flags_and_payloads('distinct_id_of_your_user')
             ```
 
         Category:
@@ -2062,12 +2133,17 @@ class Client(object):
             )
         )
 
+        # Resolve device_id from context if not provided
+        if device_id is None:
+            device_id = get_context_device_id()
+
         response, fallback_to_flags = self._get_all_flags_and_payloads_locally(
             distinct_id,
             groups=groups,
             person_properties=person_properties,
             group_properties=group_properties,
             flag_keys_to_evaluate=flag_keys_to_evaluate,
+            device_id=device_id,
         )
 
         if fallback_to_flags and not only_evaluate_locally:
@@ -2079,6 +2155,7 @@ class Client(object):
                     group_properties=group_properties,
                     disable_geoip=disable_geoip,
                     flag_keys_to_evaluate=flag_keys_to_evaluate,
+                    device_id=device_id,
                 )
                 return to_flags_and_payloads(decide_response)
             except Exception as e:
@@ -2097,6 +2174,7 @@ class Client(object):
         group_properties=None,
         warn_on_unknown_groups=False,
         flag_keys_to_evaluate: Optional[list[str]] = None,
+        device_id: Optional[str] = None,
     ) -> tuple[FlagsAndPayloads, bool]:
         person_properties = person_properties or {}
         group_properties = group_properties or {}
@@ -2126,6 +2204,7 @@ class Client(object):
                         person_properties=person_properties,
                         group_properties=group_properties,
                         warn_on_unknown_groups=warn_on_unknown_groups,
+                        device_id=device_id,
                     )
                     matched_payload = self._compute_payload_locally(
                         flag["key"], flags[flag["key"]]
@@ -2152,7 +2231,7 @@ class Client(object):
         """Initialize feature flag cache for graceful degradation during service outages.
 
         When enabled, the cache stores flag evaluation results and serves them as fallback
-        when the PostHog API is unavailable. This ensures your application continues to
+        when the Insights API is unavailable. This ensures your application continues to
         receive flag values even during outages.
 
         Args:

hanzo_insights/consumer.py

@@ -3,9 +3,7 @@ import logging
 import time
 from threading import Thread
 
-import backoff
-
-from posthog.request import APIError, DatetimeSerializer, batch_post
+from hanzo_insights.request import APIError, DatetimeSerializer, batch_post
 
 try:
     from queue import Empty
@@ -23,7 +21,7 @@ BATCH_SIZE_LIMIT = 5 * 1024 * 1024
 class Consumer(Thread):
     """Consumes the messages from the client's queue."""
 
-    log = logging.getLogger("posthog")
+    log = logging.getLogger("hanzo_insights")
 
     def __init__(
         self,
@@ -84,12 +82,16 @@ class Consumer(Thread):
             self.log.error("error uploading: %s", e)
             success = False
             if self.on_error:
-                self.on_error(e, batch)
+                try:
+                    self.on_error(e, batch)
+                except Exception as e:
+                    self.log.error("on_error handler failed: %s", e)
         finally:
             # mark items as acknowledged from queue
             for item in batch:
                 self.queue.task_done()
-            return success
+
+        return success
 
     def next(self):
         """Return the next batch of items to upload."""
@@ -124,29 +126,41 @@ class Consumer(Thread):
     def request(self, batch):
         """Attempt to upload the batch and retry before raising an error"""
 
-        def fatal_exception(exc):
+        def is_retryable(exc):
             if isinstance(exc, APIError):
                 # retry on server errors and client errors
-                # with 429 status code (rate limited),
+                # with 408 (request timeout) or 429 (rate limited),
                 # don't retry on other client errors
                 if exc.status == "N/A":
                     return False
-                return (400 <= exc.status < 500) and exc.status != 429
+                return not ((400 <= exc.status < 500) and exc.status not in (408, 429))
             else:
                 # retry on all other errors (eg. network)
-                return False
+                return True
 
-        @backoff.on_exception(
-            backoff.expo, Exception, max_tries=self.retries + 1, giveup=fatal_exception
-        )
-        def send_request():
-            batch_post(
-                self.api_key,
-                self.host,
-                gzip=self.gzip,
-                timeout=self.timeout,
-                batch=batch,
-                historical_migration=self.historical_migration,
-            )
+        last_exc = None
+        for attempt in range(self.retries + 1):
+            try:
+                batch_post(
+                    self.api_key,
+                    self.host,
+                    gzip=self.gzip,
+                    timeout=self.timeout,
+                    batch=batch,
+                    historical_migration=self.historical_migration,
+                )
+                return
+            except Exception as e:
+                last_exc = e
+                if not is_retryable(e):
+                    raise
+                if attempt < self.retries:
+                    # Respect Retry-After header if present, otherwise use exponential backoff
+                    retry_after = getattr(e, "retry_after", None)
+                    if retry_after and retry_after > 0:
+                        time.sleep(retry_after)
+                    else:
+                        time.sleep(min(2**attempt, 30))
 
-        send_request()
+        if last_exc:
+            raise last_exc

hanzo_insights/contexts.py

@@ -4,7 +4,7 @@ from typing import Optional, Any, Callable, Dict, TypeVar, cast, TYPE_CHECKING
 
 if TYPE_CHECKING:
     # To avoid circular imports
-    from posthog.client import Client
+    from hanzo_insights.client import Client
 
 
 class ContextScope:
@@ -21,6 +21,7 @@ class ContextScope:
         self.capture_exceptions = capture_exceptions
         self.session_id: Optional[str] = None
         self.distinct_id: Optional[str] = None
+        self.device_id: Optional[str] = None
         self.tags: Dict[str, Any] = {}
         self.capture_exception_code_variables: Optional[bool] = None
         self.code_variables_mask_patterns: Optional[list] = None
@@ -32,6 +33,9 @@ class ContextScope:
     def set_distinct_id(self, distinct_id: str):
         self.distinct_id = distinct_id
 
+    def set_device_id(self, device_id: str):
+        self.device_id = device_id
+
     def add_tag(self, key: str, value: Any):
         self.tags[key] = value
 
@@ -61,15 +65,21 @@ class ContextScope:
             return self.parent.get_distinct_id()
         return None
 
+    def get_device_id(self) -> Optional[str]:
+        if self.device_id is not None:
+            return self.device_id
+        if self.parent is not None and not self.fresh:
+            return self.parent.get_device_id()
+        return None
+
     def collect_tags(self) -> Dict[str, Any]:
-        tags = self.tags.copy()
         if self.parent and not self.fresh:
             # We want child tags to take precedence over parent tags,
-            # so we can't use a simple update here, instead collecting
-            # the parent tags and then updating with the child tags.
-            new_tags = self.parent.collect_tags()
-            tags.update(new_tags)
-        return tags
+            # so collect parent tags first, then update with child tags.
+            tags = self.parent.collect_tags()
+            tags.update(self.tags)
+            return tags
+        return self.tags.copy()
 
     def get_capture_exception_code_variables(self) -> Optional[bool]:
         if self.capture_exception_code_variables is not None:
@@ -94,7 +104,7 @@ class ContextScope:
 
 
 _context_stack: contextvars.ContextVar[Optional[ContextScope]] = contextvars.ContextVar(
-    "posthog_context_stack", default=None
+    "insights_context_stack", default=None
 )
 
 
@@ -124,32 +134,32 @@ def new_context(
                 If provided, the client will be used to capture exceptions within the context.
                 If not provided, the default (global) client will be used. Note that the passed
                 client is only used to capture exceptions within the context - other events captured
-                within the context via `Client.capture` or `posthog.capture` will still carry the context
+                within the context via `Client.capture` or `hanzo_insights.capture` will still carry the context
                 state (tags, identity, session id), but will be captured by the client directly used (or
-                the global one, in the case of `posthog.capture`)
+                the global one, in the case of `hanzo_insights.capture`)
 
     Examples:
         ```python
         # Inherit parent context tags
-        with posthog.new_context():
-            posthog.tag("request_id", "123")
+        with hanzo_insights.new_context():
+            hanzo_insights.tag("request_id", "123")
             # Both this event and the exception will be tagged with the context tags
-            posthog.capture("event_name", {"property": "value"})
+            hanzo_insights.capture("event_name", {"property": "value"})
             raise ValueError("Something went wrong")
         ```
         ```python
         # Start with fresh context (no inherited tags)
-        with posthog.new_context(fresh=True):
-            posthog.tag("request_id", "123")
+        with hanzo_insights.new_context(fresh=True):
+            hanzo_insights.tag("request_id", "123")
             # Both this event and the exception will be tagged with the context tags
-            posthog.capture("event_name", {"property": "value"})
+            hanzo_insights.capture("event_name", {"property": "value"})
             raise ValueError("Something went wrong")
         ```
 
     Category:
         Contexts
     """
-    from posthog import capture_exception
+    from hanzo_insights import capture_exception
 
     current_context = _get_current_context()
     new_context = ContextScope(current_context, fresh, capture_exceptions, client)
@@ -179,7 +189,7 @@ def tag(key: str, value: Any) -> None:
 
     Example:
         ```python
-        posthog.tag("user_id", "123")
+        hanzo_insights.tag("user_id", "123")
         ```
 
     Category:
@@ -211,7 +221,7 @@ def identify_context(distinct_id: str) -> None:
     """
     Identify the current context with a distinct ID, associating all events captured in this or
     child contexts with the given distinct ID (unless identify_context is called again). This is overridden by
-    distinct id's passed directly to posthog.capture and related methods (identify, set etc). Entering a
+    distinct id's passed directly to hanzo_insights.capture and related methods (identify, set etc). Entering a
     fresh context will clear the context-level distinct ID. The distinct-id passed should be uniquely associated
     with one of your users. Events captured outside of a context, or in a context with no associated distinct
     ID, will be assigned a random UUID, and captured as "personless".
@@ -234,7 +244,7 @@ def set_context_session(session_id: str) -> None:
     Entering a fresh context will clear the context-level session ID.
 
     Args:
-        session_id: The session ID to associate with the current context and its children. See https://posthog.com/docs/data/sessions
+        session_id: The session ID to associate with the current context and its children. See https://insights.hanzo.ai/docs/data/sessions
 
     Category:
         Contexts
@@ -276,6 +286,39 @@ def get_context_distinct_id() -> Optional[str]:
     return None
 
 
+def set_context_device_id(device_id: str) -> None:
+    """
+    Set the device ID for the current context, associating all feature flag requests in this or
+    child contexts with the given device ID (unless set_context_device_id is called again).
+    Entering a fresh context will clear the context-level device ID.
+
+    Args:
+        device_id: The device ID to associate with the current context and its children.
+
+    Category:
+        Contexts
+    """
+    current_context = _get_current_context()
+    if current_context:
+        current_context.set_device_id(device_id)
+
+
+def get_context_device_id() -> Optional[str]:
+    """
+    Get the device ID for the current context.
+
+    Returns:
+        The device ID if set, None otherwise
+
+    Category:
+        Contexts
+    """
+    current_context = _get_current_context()
+    if current_context:
+        return current_context.get_device_id()
+    return None
+
+
 def set_capture_exception_code_variables_context(enabled: bool) -> None:
     """
     Set whether code variables are captured for the current context.
@@ -330,20 +373,20 @@ F = TypeVar("F", bound=Callable[..., Any])
 def scoped(fresh: bool = False, capture_exceptions: bool = True):
     """
     Decorator that creates a new context for the function. Simply wraps
-    the function in a with posthog.new_context(): block.
+    the function in a with hanzo_insights.new_context(): block.
 
     Args:
         fresh: Whether to start with a fresh context (default: False)
-        capture_exceptions: Whether to capture and track exceptions with posthog error tracking (default: True)
+        capture_exceptions: Whether to capture and track exceptions with Insights error tracking (default: True)
 
     Example:
-        @posthog.scoped()
+        @hanzo_insights.scoped()
         def process_payment(payment_id):
-            posthog.tag("payment_id", payment_id)
-            posthog.tag("payment_method", "credit_card")
+            hanzo_insights.tag("payment_id", payment_id)
+            hanzo_insights.tag("payment_method", "credit_card")
 
             # This event will be captured with tags
-            posthog.capture("payment_started")
+            hanzo_insights.capture("payment_started")
             # If this raises an exception, it will be captured with tags
             # and then re-raised
             some_risky_function()

hanzo_insights/exception_capture.py

@@ -9,13 +9,13 @@ import threading
 from typing import TYPE_CHECKING
 
 if TYPE_CHECKING:
-    from posthog.client import Client
+    from hanzo_insights.client import Client
 
 
 class ExceptionCapture:
     # TODO: Add client side rate limiting to prevent spamming the server with exceptions
 
-    log = logging.getLogger("posthog")
+    log = logging.getLogger("hanzo_insights")
 
     def __init__(self, client: "Client"):
         self.client = client

hanzo_insights/exception_utils.py

@@ -30,7 +30,7 @@ from typing import (  # noqa: F401
     cast,
 )
 
-from posthog.args import ExceptionArg, ExcInfo  # noqa: F401
+from hanzo_insights.args import ExceptionArg, ExcInfo  # noqa: F401
 
 try:
     # Python 3.11
@@ -43,26 +43,31 @@ except ImportError:
 DEFAULT_MAX_VALUE_LENGTH = 1024
 
 DEFAULT_CODE_VARIABLES_MASK_PATTERNS = [
-    r"(?i).*password.*",
-    r"(?i).*secret.*",
-    r"(?i).*passwd.*",
-    r"(?i).*pwd.*",
-    r"(?i).*api_key.*",
-    r"(?i).*apikey.*",
-    r"(?i).*auth.*",
-    r"(?i).*credentials.*",
-    r"(?i).*privatekey.*",
-    r"(?i).*private_key.*",
-    r"(?i).*token.*",
-    r"(?i).*aws_access_key_id.*",
-    r"(?i).*_pass",
-    r"(?i)sk_.*",
-    r"(?i).*jwt.*",
+    r"(?i)password",
+    r"(?i)secret",
+    r"(?i)passwd",
+    r"(?i)pwd",
+    r"(?i)api_key",
+    r"(?i)apikey",
+    r"(?i)auth",
+    r"(?i)credentials",
+    r"(?i)privatekey",
+    r"(?i)private_key",
+    r"(?i)token",
+    r"(?i)aws_access_key_id",
+    r"(?i)_pass",
+    r"(?i)sk_",
+    r"(?i)jwt",
 ]
 
 DEFAULT_CODE_VARIABLES_IGNORE_PATTERNS = [r"^__.*"]
 
-CODE_VARIABLES_REDACTED_VALUE = "$$_posthog_redacted_based_on_masking_rules_$$"
+CODE_VARIABLES_REDACTED_VALUE = "$$_insights_redacted_based_on_masking_rules_$$"
+CODE_VARIABLES_TOO_LONG_VALUE = "$$_insights_value_too_long_$$"
+
+_MAX_VALUE_LENGTH_FOR_PATTERN_MATCH = 5_000
+_MAX_COLLECTION_ITEMS_TO_SCAN = 100
+_REGEX_METACHARACTERS = frozenset(r"\.^$*+?{}[]|()")
 
 DEFAULT_TOTAL_VARIABLES_SIZE_LIMIT = 20 * 1024
 
@@ -763,12 +768,12 @@ def set_in_app_in_frames(frames, in_app_exclude, in_app_include, project_root=No
 def exception_is_already_captured(error):
     # type: (ExceptionArg) -> bool
     if isinstance(error, BaseException):
-        return hasattr(error, "__posthog_exception_captured")
+        return hasattr(error, "__insights_exception_captured")
     # Autocaptured exceptions are passed as a tuple from our system hooks,
     # the second item is the exception value (the first is the exception type)
     elif isinstance(error, tuple) and len(error) > 1:
         return error[1] is not None and hasattr(
-            error[1], "__posthog_exception_captured"
+            error[1], "__insights_exception_captured"
         )
     else:
         return False  # type: ignore[unreachable]
@@ -777,14 +782,14 @@ def exception_is_already_captured(error):
 def mark_exception_as_captured(error, uuid):
     # type: (ExceptionArg, str) -> None
     if isinstance(error, BaseException):
-        setattr(error, "__posthog_exception_captured", True)
-        setattr(error, "__posthog_exception_uuid", uuid)
+        setattr(error, "__insights_exception_captured", True)
+        setattr(error, "__insights_exception_uuid", uuid)
     # Autocaptured exceptions are passed as a tuple from our system hooks,
     # the second item is the exception value (the first is the exception type)
     elif isinstance(error, tuple) and len(error) > 1:
         if error[1] is not None:
-            setattr(error[1], "__posthog_exception_captured", True)
-            setattr(error[1], "__posthog_exception_uuid", uuid)
+            setattr(error[1], "__insights_exception_captured", True)
+            setattr(error[1], "__insights_exception_uuid", uuid)
 
 
 def exc_info_from_error(error):
@@ -928,40 +933,87 @@ def strip_string(value, max_length=None):
     )
 
 
+def _extract_plain_substring(pattern):
+    # Matches inline flag groups like (?i), (?ai), (?ims), etc. that include the 'i' flag.
+    # Python regex flags: a=ASCII, i=IGNORECASE, L=LOCALE, m=MULTILINE, s=DOTALL, u=UNICODE, x=VERBOSE
+    inline_flags = re.match(r"^\(\?[aiLmsux]*i[aiLmsux]*\)", pattern)
+    if not inline_flags:
+        return None
+    remainder = pattern[inline_flags.end() :]
+    if not remainder or any(c in _REGEX_METACHARACTERS for c in remainder):
+        return None
+    return remainder.lower()
+
+
 def _compile_patterns(patterns):
-    compiled = []
+    if not patterns:
+        return None
+    substrings = []
+    regexes = []
     for pattern in patterns:
-        try:
-            compiled.append(re.compile(pattern))
-        except Exception:
-            pass
-    return compiled
+        simple = _extract_plain_substring(pattern)
+        if simple is not None:
+            substrings.append(simple)
+        else:
+            try:
+                regexes.append(re.compile(pattern))
+            except Exception:
+                pass
+    if not substrings and not regexes:
+        return None
+    return (substrings, regexes)
 
 
 def _pattern_matches(name, patterns):
-    for pattern in patterns:
+    if patterns is None:
+        return False
+    substrings, regexes = patterns
+    if substrings:
+        name_lower = name.lower()
+        for s in substrings:
+            if s in name_lower:
+                return True
+    for pattern in regexes:
         if pattern.search(name):
             return True
     return False
 
 
-def _mask_sensitive_data(value, compiled_mask):
+def _mask_sensitive_data(value, compiled_mask, _seen=None):
     if not compiled_mask:
         return value
 
+    if isinstance(value, (dict, list, tuple)):
+        if _seen is None:
+            _seen = set()
+        obj_id = id(value)
+        if obj_id in _seen:
+            return "<circular ref>"
+        _seen.add(obj_id)
+
     if isinstance(value, dict):
+        if len(value) > _MAX_COLLECTION_ITEMS_TO_SCAN:
+            return CODE_VARIABLES_TOO_LONG_VALUE
         result = {}
         for k, v in value.items():
             key_str = str(k) if not isinstance(k, str) else k
-            if _pattern_matches(key_str, compiled_mask):
+            if len(key_str) > _MAX_VALUE_LENGTH_FOR_PATTERN_MATCH:
+                result[k] = CODE_VARIABLES_TOO_LONG_VALUE
+            elif _pattern_matches(key_str, compiled_mask):
                 result[k] = CODE_VARIABLES_REDACTED_VALUE
             else:
-                result[k] = _mask_sensitive_data(v, compiled_mask)
+                result[k] = _mask_sensitive_data(v, compiled_mask, _seen)
         return result
     elif isinstance(value, (list, tuple)):
-        masked_items = [_mask_sensitive_data(item, compiled_mask) for item in value]
+        if len(value) > _MAX_COLLECTION_ITEMS_TO_SCAN:
+            return CODE_VARIABLES_TOO_LONG_VALUE
+        masked_items = [
+            _mask_sensitive_data(item, compiled_mask, _seen) for item in value
+        ]
         return type(value)(masked_items)
     elif isinstance(value, str):
+        if len(value) > _MAX_VALUE_LENGTH_FOR_PATTERN_MATCH:
+            return CODE_VARIABLES_TOO_LONG_VALUE
         if _pattern_matches(value, compiled_mask):
             return CODE_VARIABLES_REDACTED_VALUE
         return value
@@ -982,7 +1034,9 @@ def _serialize_variable_value(value, limiter, max_length=1024, compiled_mask=Non
             limiter.add(result_size)
             return value
         elif isinstance(value, str):
-            if compiled_mask and _pattern_matches(value, compiled_mask):
+            if len(value) > _MAX_VALUE_LENGTH_FOR_PATTERN_MATCH:
+                result = CODE_VARIABLES_TOO_LONG_VALUE
+            elif compiled_mask and _pattern_matches(value, compiled_mask):
                 result = CODE_VARIABLES_REDACTED_VALUE
             else:
                 result = value

hanzo_insights/feature_flags.py

@@ -2,21 +2,46 @@ import datetime
 import hashlib
 import logging
 import re
+import warnings
 from typing import Optional
 
 from dateutil import parser
 from dateutil.relativedelta import relativedelta
 
-from posthog import utils
-from posthog.types import FlagValue
-from posthog.utils import convert_to_datetime_aware, is_valid_regex
+from hanzo_insights import utils
+from hanzo_insights.types import FlagValue
+from hanzo_insights.utils import convert_to_datetime_aware, is_valid_regex
 
 __LONG_SCALE__ = float(0xFFFFFFFFFFFFFFF)
 
-log = logging.getLogger("posthog")
+log = logging.getLogger("hanzo_insights")
 
 NONE_VALUES_ALLOWED_OPERATORS = ["is_not"]
 
+# All operators supported by match_property, grouped by category.
+EQUALITY_OPERATORS = ("exact", "is_not", "is_set", "is_not_set")
+STRING_OPERATORS = ("icontains", "not_icontains", "regex", "not_regex")
+NUMERIC_OPERATORS = ("gt", "gte", "lt", "lte")
+DATE_OPERATORS = ("is_date_before", "is_date_after")
+SEMVER_COMPARISON_OPERATORS = (
+    "semver_eq",
+    "semver_neq",
+    "semver_gt",
+    "semver_gte",
+    "semver_lt",
+    "semver_lte",
+)
+SEMVER_RANGE_OPERATORS = ("semver_tilde", "semver_caret", "semver_wildcard")
+SEMVER_OPERATORS = SEMVER_COMPARISON_OPERATORS + SEMVER_RANGE_OPERATORS
+
+PROPERTY_OPERATORS = (
+    EQUALITY_OPERATORS
+    + STRING_OPERATORS
+    + NUMERIC_OPERATORS
+    + DATE_OPERATORS
+    + SEMVER_OPERATORS
+)
+
 
 class InconclusiveMatchError(Exception):
     pass
@@ -34,18 +59,18 @@ class RequiresServerEvaluation(Exception):
     pass
 
 
-# This function takes a distinct_id and a feature flag key and returns a float between 0 and 1.
-# Given the same distinct_id and key, it'll always return the same float. These floats are
+# This function takes a bucketing value and a feature flag key and returns a float between 0 and 1.
+# Given the same bucketing value and key, it'll always return the same float. These floats are
 # uniformly distributed between 0 and 1, so if we want to show this feature to 20% of traffic
-# we can do _hash(key, distinct_id) < 0.2
-def _hash(key: str, distinct_id: str, salt: str = "") -> float:
-    hash_key = f"{key}.{distinct_id}{salt}"
+# we can do _hash(key, bucketing_value) < 0.2
+def _hash(key: str, bucketing_value: str, salt: str = "") -> float:
+    hash_key = f"{key}.{bucketing_value}{salt}"
     hash_val = int(hashlib.sha1(hash_key.encode("utf-8")).hexdigest()[:15], 16)
     return hash_val / __LONG_SCALE__
 
 
-def get_matching_variant(flag, distinct_id):
-    hash_value = _hash(flag["key"], distinct_id, salt="variant")
+def get_matching_variant(flag, bucketing_value):
+    hash_value = _hash(flag["key"], bucketing_value, salt="variant")
     for variant in variant_lookup_table(flag):
         if hash_value >= variant["value_min"] and hash_value < variant["value_max"]:
             return variant["key"]
@@ -68,7 +93,13 @@ def variant_lookup_table(feature_flag):
 
 
 def evaluate_flag_dependency(
-    property, flags_by_key, evaluation_cache, distinct_id, properties, cohort_properties
+    property,
+    flags_by_key,
+    evaluation_cache,
+    distinct_id,
+    properties,
+    cohort_properties,
+    device_id=None,
 ):
     """
     Evaluate a flag dependency property according to the dependency chain algorithm.
@@ -80,6 +111,7 @@ def evaluate_flag_dependency(
         distinct_id: The distinct ID being evaluated
         properties: Person properties for evaluation
         cohort_properties: Cohort properties for evaluation
+        device_id: The device ID for bucketing (optional)
 
     Returns:
         bool: True if all dependencies in the chain evaluate to True, False otherwise
@@ -124,13 +156,27 @@ def evaluate_flag_dependency(
                 else:
                     # Recursively evaluate the dependency
                     try:
+                        dep_flag_filters = dep_flag.get("filters") or {}
+                        dep_aggregation_group_type_index = dep_flag_filters.get(
+                            "aggregation_group_type_index"
+                        )
+                        if dep_aggregation_group_type_index is not None:
+                            # Group flags should continue bucketing by the group key
+                            # from the current evaluation context.
+                            dep_bucketing_value = distinct_id
+                        else:
+                            dep_bucketing_value = resolve_bucketing_value(
+                                dep_flag, distinct_id, device_id
+                            )
                         dep_result = match_feature_flag_properties(
                             dep_flag,
                             distinct_id,
                             properties,
-                            cohort_properties,
-                            flags_by_key,
-                            evaluation_cache,
+                            cohort_properties=cohort_properties,
+                            flags_by_key=flags_by_key,
+                            evaluation_cache=evaluation_cache,
+                            device_id=device_id,
+                            bucketing_value=dep_bucketing_value,
                         )
                         evaluation_cache[dep_flag_key] = dep_result
                     except InconclusiveMatchError as e:
@@ -215,21 +261,54 @@ def matches_dependency_value(expected_value, actual_value):
     return False
 
 
+def resolve_bucketing_value(flag, distinct_id, device_id=None):
+    """Resolve the bucketing value for a flag based on its bucketing_identifier setting.
+
+    Returns:
+        The appropriate identifier string to use for hashing/bucketing.
+
+    Raises:
+        InconclusiveMatchError: If the flag requires device_id but none was provided.
+    """
+    flag_filters = flag.get("filters") or {}
+    bucketing_identifier = flag.get("bucketing_identifier") or flag_filters.get(
+        "bucketing_identifier"
+    )
+    if bucketing_identifier == "device_id":
+        if not device_id:
+            raise InconclusiveMatchError(
+                "Flag requires device_id for bucketing but none was provided"
+            )
+        return device_id
+    return distinct_id
+
+
 def match_feature_flag_properties(
     flag,
     distinct_id,
     properties,
+    *,
     cohort_properties=None,
     flags_by_key=None,
     evaluation_cache=None,
+    device_id=None,
+    bucketing_value=None,
 ) -> FlagValue:
-    flag_conditions = (flag.get("filters") or {}).get("groups") or []
+    if bucketing_value is None:
+        warnings.warn(
+            "Calling match_feature_flag_properties() without bucketing_value is deprecated. "
+            "Pass bucketing_value explicitly. This fallback will be removed in a future major release.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+        bucketing_value = resolve_bucketing_value(flag, distinct_id, device_id)
+
+    flag_filters = flag.get("filters") or {}
+    flag_conditions = flag_filters.get("groups") or []
     is_inconclusive = False
     cohort_properties = cohort_properties or {}
     # Some filters can be explicitly set to null, which require accessing variants like so
-    flag_variants = ((flag.get("filters") or {}).get("multivariate") or {}).get(
-        "variants"
-    ) or []
+    flag_variants = (flag_filters.get("multivariate") or {}).get("variants") or []
     valid_variant_keys = [variant["key"] for variant in flag_variants]
 
     for condition in flag_conditions:
@@ -244,12 +323,14 @@ def match_feature_flag_properties(
                 cohort_properties,
                 flags_by_key,
                 evaluation_cache,
+                bucketing_value=bucketing_value,
+                device_id=device_id,
             ):
                 variant_override = condition.get("variant")
                 if variant_override and variant_override in valid_variant_keys:
                     variant = variant_override
                 else:
-                    variant = get_matching_variant(flag, distinct_id)
+                    variant = get_matching_variant(flag, bucketing_value)
                 return variant or True
         except RequiresServerEvaluation:
             # Static cohort or other missing server-side data - must fallback to API
@@ -277,6 +358,9 @@ def is_condition_match(
     cohort_properties,
     flags_by_key=None,
     evaluation_cache=None,
+    *,
+    bucketing_value,
+    device_id=None,
 ) -> bool:
     rollout_percentage = condition.get("rollout_percentage")
     if len(condition.get("properties") or []) > 0:
@@ -290,6 +374,7 @@ def is_condition_match(
                     flags_by_key,
                     evaluation_cache,
                     distinct_id,
+                    device_id=device_id,
                 )
             elif property_type == "flag":
                 matches = evaluate_flag_dependency(
@@ -299,6 +384,7 @@ def is_condition_match(
                     distinct_id,
                     properties,
                     cohort_properties,
+                    device_id=device_id,
                 )
             else:
                 matches = match_property(prop, properties)
@@ -308,9 +394,9 @@ def is_condition_match(
         if rollout_percentage is None:
             return True
 
-    if rollout_percentage is not None and _hash(feature_flag["key"], distinct_id) > (
-        rollout_percentage / 100
-    ):
+    if rollout_percentage is not None and _hash(
+        feature_flag["key"], bucketing_value
+    ) > (rollout_percentage / 100):
         return False
 
     return True
@@ -323,6 +409,9 @@ def match_property(property, property_values) -> bool:
     operator = property.get("operator") or "exact"
     value = property.get("value")
 
+    if operator not in PROPERTY_OPERATORS:
+        raise InconclusiveMatchError(f"Unknown operator {operator}")
+
     if key not in property_values:
         raise InconclusiveMatchError(
             "can't match properties without a given property value"
@@ -443,7 +532,64 @@ def match_property(property, property_values) -> bool:
                 "The date provided must be a string or date object"
             )
 
-    # if we get here, we don't know how to handle the operator
+    if operator in SEMVER_OPERATORS:
+        try:
+            override_parsed = parse_semver(override_value)
+        except (ValueError, TypeError):
+            raise InconclusiveMatchError(
+                f"Person property value '{override_value}' is not a valid semver"
+            )
+
+        if operator in SEMVER_COMPARISON_OPERATORS:
+            try:
+                flag_parsed = parse_semver(value)
+            except (ValueError, TypeError):
+                raise InconclusiveMatchError(
+                    f"Flag semver value '{value}' is not a valid semver"
+                )
+
+            if operator == "semver_eq":
+                return override_parsed == flag_parsed
+            elif operator == "semver_neq":
+                return override_parsed != flag_parsed
+            elif operator == "semver_gt":
+                return override_parsed > flag_parsed
+            elif operator == "semver_gte":
+                return override_parsed >= flag_parsed
+            elif operator == "semver_lt":
+                return override_parsed < flag_parsed
+            elif operator == "semver_lte":
+                return override_parsed <= flag_parsed
+
+        elif operator == "semver_tilde":
+            try:
+                lower, upper = _tilde_bounds(str(value))
+            except (ValueError, TypeError):
+                raise InconclusiveMatchError(
+                    f"Flag semver value '{value}' is not valid for tilde operator"
+                )
+            return lower <= override_parsed < upper
+
+        elif operator == "semver_caret":
+            try:
+                lower, upper = _caret_bounds(str(value))
+            except (ValueError, TypeError):
+                raise InconclusiveMatchError(
+                    f"Flag semver value '{value}' is not valid for caret operator"
+                )
+            return lower <= override_parsed < upper
+
+        elif operator == "semver_wildcard":
+            try:
+                lower, upper = _wildcard_bounds(str(value))
+            except (ValueError, TypeError):
+                raise InconclusiveMatchError(
+                    f"Flag semver value '{value}' is not valid for wildcard operator"
+                )
+            return lower <= override_parsed < upper
+
+    # Unreachable: all operators in PROPERTY_OPERATORS are handled above,
+    # and unknown operators are rejected at the top of this function.
     raise InconclusiveMatchError(f"Unknown operator {operator}")
 
 
@@ -454,6 +600,7 @@ def match_cohort(
     flags_by_key=None,
     evaluation_cache=None,
     distinct_id=None,
+    device_id=None,
 ) -> bool:
     # Cohort properties are in the form of property groups like this:
     # {
@@ -478,6 +625,7 @@ def match_cohort(
         flags_by_key,
         evaluation_cache,
         distinct_id,
+        device_id=device_id,
     )
 
 
@@ -488,6 +636,7 @@ def match_property_group(
     flags_by_key=None,
     evaluation_cache=None,
     distinct_id=None,
+    device_id=None,
 ) -> bool:
     if not property_group:
         return True
@@ -512,6 +661,7 @@ def match_property_group(
                     flags_by_key,
                     evaluation_cache,
                     distinct_id,
+                    device_id=device_id,
                 )
                 if property_group_type == "AND":
                     if not matches:
@@ -545,6 +695,7 @@ def match_property_group(
                         flags_by_key,
                         evaluation_cache,
                         distinct_id,
+                        device_id=device_id,
                     )
                 elif prop.get("type") == "flag":
                     matches = evaluate_flag_dependency(
@@ -554,6 +705,7 @@ def match_property_group(
                         distinct_id,
                         property_values,
                         cohort_properties,
+                        device_id=device_id,
                     )
                 else:
                     matches = match_property(prop, property_values)
@@ -618,3 +770,75 @@ def relative_date_parse_for_feature_flag_matching(
         return parsed_dt
     else:
         return None
+
+
+def parse_semver(value: str) -> tuple:
+    """Parse a semver string into a comparable (major, minor, patch) integer tuple.
+
+    Matches the behavior of the sortableSemver HogQL function:
+    - Handles v-prefix, whitespace, pre-release suffixes
+    - Defaults missing components to 0 (e.g., 1.2 -> 1.2.0)
+    Raises ValueError if parsing fails.
+    """
+    text = str(value).strip().lstrip("vV")
+    # Strip pre-release/build metadata suffix
+    text = text.split("-")[0].split("+")[0]
+    parts = text.split(".")
+
+    if not parts or not parts[0]:
+        raise ValueError("Invalid semver format")
+
+    major = int(parts[0])
+    minor = int(parts[1]) if len(parts) > 1 and parts[1] else 0
+    patch = int(parts[2]) if len(parts) > 2 and parts[2] else 0
+
+    return (major, minor, patch)
+
+
+def _tilde_bounds(value: str) -> tuple:
+    """~1.2.3 means >=1.2.3 <1.3.0 (allows patch-level changes)."""
+    major, minor, patch = parse_semver(value)
+    return (major, minor, patch), (major, minor + 1, 0)
+
+
+def _caret_bounds(value: str) -> tuple:
+    """Caret follows semver spec:
+    ^1.2.3 means >=1.2.3 <2.0.0
+    ^0.2.3 means >=0.2.3 <0.3.0
+    ^0.0.3 means >=0.0.3 <0.0.4
+    """
+    major, minor, patch = parse_semver(value)
+    lower = (major, minor, patch)
+
+    if major > 0:
+        upper = (major + 1, 0, 0)
+    elif minor > 0:
+        upper = (0, minor + 1, 0)
+    else:
+        upper = (0, 0, patch + 1)
+
+    return lower, upper
+
+
+def _wildcard_bounds(value: str) -> tuple:
+    """Wildcard matching:
+    1.* means >=1.0.0 <2.0.0
+    1.2.* means >=1.2.0 <1.3.0
+    """
+    cleaned = str(value).strip().lstrip("vV").replace("*", "").rstrip(".")
+    if not cleaned:
+        raise ValueError("Invalid wildcard pattern")
+
+    parts = [p for p in cleaned.split(".") if p]
+    if not parts:
+        raise ValueError("Invalid wildcard pattern")
+
+    if len(parts) == 1:
+        major = int(parts[0])
+        return (major, 0, 0), (major + 1, 0, 0)
+    elif len(parts) == 2:
+        major, minor = int(parts[0]), int(parts[1])
+        return (major, minor, 0), (major, minor + 1, 0)
+    else:
+        major, minor, patch = int(parts[0]), int(parts[1]), int(parts[2])
+        return (major, minor, patch), (major, minor, patch + 1)

hanzo_insights/flag_definition_cache.py

@@ -9,11 +9,11 @@ functions) to share flag definitions and reduce API calls.
 
 Usage:
 
-    from posthog import Posthog
-    from posthog.flag_definition_cache import FlagDefinitionCacheProvider
+    from hanzo_insights import Insights
+    from hanzo_insights.flag_definition_cache import FlagDefinitionCacheProvider
 
     cache = RedisFlagDefinitionCache(redis_client, "my-team")
-    posthog = Posthog(
+    client = Insights(
         "<project_api_key>",
         personal_api_key="<personal_api_key>",
         flag_definition_cache_provider=cache,
@@ -63,7 +63,7 @@ class FlagDefinitionCacheProvider(Protocol):
        new definitions from the API. Store the data in your external cache
        and release any locks.
 
-    4. `shutdown()` - Called when the PostHog client shuts down. Release any
+    4. `shutdown()` - Called when the Insights client shuts down. Release any
        distributed locks and clean up resources.
 
     Error Handling:
@@ -104,7 +104,7 @@ class FlagDefinitionCacheProvider(Protocol):
 
     def on_flag_definitions_received(self, data: FlagDefinitionCacheData) -> None:
         """
-        Called after successfully receiving new flag definitions from PostHog.
+        Called after successfully receiving new flag definitions from Insights.
 
         Use this to store the data in your external cache and release any
         distributed locks acquired in `should_fetch_flag_definitions()`.
@@ -117,7 +117,7 @@ class FlagDefinitionCacheProvider(Protocol):
 
     def shutdown(self) -> None:
         """
-        Called when the PostHog client shuts down.
+        Called when the Insights client shuts down.
 
         Use this to release any distributed locks and clean up resources.
         This method is called even if `should_fetch_flag_definitions()`

hanzo_insights/integrations/django.py

@@ -1,6 +1,6 @@
 from typing import TYPE_CHECKING, cast
-from posthog import contexts
-from posthog.client import Client
+from hanzo_insights import contexts
+from hanzo_insights.client import Client
 
 try:
     from asgiref.sync import iscoroutinefunction, markcoroutinefunction
@@ -21,25 +21,26 @@ if TYPE_CHECKING:
     from typing import Callable, Dict, Any, Optional, Union, Awaitable  # noqa: F401
 
 
-class PosthogContextMiddleware:
+class InsightsContextMiddleware:
     """Middleware to automatically track Django requests.
 
-    This middleware wraps all calls with a posthog context. It attempts to extract the following from the request headers:
-    - Session ID, (extracted from `X-POSTHOG-SESSION-ID`)
-    - Distinct ID, (extracted from `X-POSTHOG-DISTINCT-ID`)
+    This middleware wraps all calls with an Insights context. It attempts to extract the following from the request headers:
+    - Session ID, (extracted from `X-INSIGHTS-SESSION-ID`)
+    - Distinct ID, (extracted from `X-INSIGHTS-DISTINCT-ID`)
     - Request URL as $current_url
     - Request Method as $request_method
 
-    The context will also auto-capture exceptions and send them to PostHog, unless you disable it by setting
-    `POSTHOG_MW_CAPTURE_EXCEPTIONS` to `False` in your Django settings. The exceptions are captured using the
-    global client, unless the setting `POSTHOG_MW_CLIENT` is set to a custom client instance
+    The context will also auto-capture exceptions and send them to Insights, unless you disable it by setting
+    `INSIGHTS_MW_CAPTURE_EXCEPTIONS` to `False` in your Django settings.
+    The exceptions are captured using the global client, unless the setting `INSIGHTS_MW_CLIENT`
+    is set to a custom client instance.
 
     The middleware behaviour is customisable through 3 additional functions:
-    - `POSTHOG_MW_EXTRA_TAGS`, which is a Callable[[HttpRequest], Dict[str, Any]] expected to return a dictionary of additional tags to be added to the context.
-    - `POSTHOG_MW_REQUEST_FILTER`, which is a Callable[[HttpRequest], bool] expected to return `False` if the request should not be tracked.
-    - `POSTHOG_MW_TAG_MAP`, which is a Callable[[Dict[str, Any]], Dict[str, Any]], which you can use to modify the tags before they're added to the context.
+    - `INSIGHTS_MW_EXTRA_TAGS`, which is a Callable[[HttpRequest], Dict[str, Any]] expected to return a dictionary of additional tags to be added to the context.
+    - `INSIGHTS_MW_REQUEST_FILTER`, which is a Callable[[HttpRequest], bool] expected to return `False` if the request should not be tracked.
+    - `INSIGHTS_MW_TAG_MAP`, which is a Callable[[Dict[str, Any]], Dict[str, Any]], which you can use to modify the tags before they're added to the context.
 
-    You can use the `POSTHOG_MW_TAG_MAP` function to remove any default tags you don't want to capture, or override them with your own values.
+    You can use the `INSIGHTS_MW_TAG_MAP` function to remove any default tags you don't want to capture, or override them with your own values.
 
     Context tags are automatically included as properties on all events captured within a context, including exceptions.
     See the context documentation for more information. The extracted distinct ID and session ID, if found, are used to
@@ -66,47 +67,48 @@ class PosthogContextMiddleware:
 
         from django.conf import settings
 
-        if hasattr(settings, "POSTHOG_MW_EXTRA_TAGS") and callable(
-            settings.POSTHOG_MW_EXTRA_TAGS
-        ):
+        def _get_setting(name):
+            insights_name = f"INSIGHTS_MW_{name}"
+            if hasattr(settings, insights_name):
+                return getattr(settings, insights_name)
+            return None
+
+        extra_tags = _get_setting("EXTRA_TAGS")
+        if extra_tags and callable(extra_tags):
             self.extra_tags = cast(
                 "Optional[Callable[[HttpRequest], Dict[str, Any]]]",
-                settings.POSTHOG_MW_EXTRA_TAGS,
+                extra_tags,
             )
         else:
             self.extra_tags = None
 
-        if hasattr(settings, "POSTHOG_MW_REQUEST_FILTER") and callable(
-            settings.POSTHOG_MW_REQUEST_FILTER
-        ):
+        request_filter = _get_setting("REQUEST_FILTER")
+        if request_filter and callable(request_filter):
             self.request_filter = cast(
                 "Optional[Callable[[HttpRequest], bool]]",
-                settings.POSTHOG_MW_REQUEST_FILTER,
+                request_filter,
             )
         else:
             self.request_filter = None
 
-        if hasattr(settings, "POSTHOG_MW_TAG_MAP") and callable(
-            settings.POSTHOG_MW_TAG_MAP
-        ):
+        tag_map = _get_setting("TAG_MAP")
+        if tag_map and callable(tag_map):
             self.tag_map = cast(
                 "Optional[Callable[[Dict[str, Any]], Dict[str, Any]]]",
-                settings.POSTHOG_MW_TAG_MAP,
+                tag_map,
             )
         else:
             self.tag_map = None
 
-        if hasattr(settings, "POSTHOG_MW_CAPTURE_EXCEPTIONS") and isinstance(
-            settings.POSTHOG_MW_CAPTURE_EXCEPTIONS, bool
-        ):
-            self.capture_exceptions = settings.POSTHOG_MW_CAPTURE_EXCEPTIONS
+        capture_exceptions = _get_setting("CAPTURE_EXCEPTIONS")
+        if isinstance(capture_exceptions, bool):
+            self.capture_exceptions = capture_exceptions
         else:
             self.capture_exceptions = True
 
-        if hasattr(settings, "POSTHOG_MW_CLIENT") and isinstance(
-            settings.POSTHOG_MW_CLIENT, Client
-        ):
-            self.client = cast("Optional[Client]", settings.POSTHOG_MW_CLIENT)
+        mw_client = _get_setting("CLIENT")
+        if isinstance(mw_client, Client):
+            self.client = cast("Optional[Client]", mw_client)
         else:
             self.client = None
 
@@ -125,13 +127,13 @@ class PosthogContextMiddleware:
         """
         tags = {}
 
-        # Extract session ID from X-POSTHOG-SESSION-ID header
-        session_id = request.headers.get("X-POSTHOG-SESSION-ID")
+        # Extract session ID from X-INSIGHTS-SESSION-ID header
+        session_id = request.headers.get("X-INSIGHTS-SESSION-ID")
         if session_id:
             contexts.set_context_session(session_id)
 
-        # Extract distinct ID from X-POSTHOG-DISTINCT-ID header or request user id
-        distinct_id = request.headers.get("X-POSTHOG-DISTINCT-ID") or user_id
+        # Extract distinct ID from X-INSIGHTS-DISTINCT-ID header or request user id
+        distinct_id = request.headers.get("X-INSIGHTS-DISTINCT-ID") or user_id
         if distinct_id:
             contexts.identify_context(distinct_id)
 
@@ -155,7 +157,7 @@ class PosthogContextMiddleware:
         # Extract IP address
         ip_address = request.headers.get("X-Forwarded-For")
         if ip_address:
-            tags["$ip_address"] = ip_address
+            tags["$ip"] = ip_address
 
         # Extract user agent
         user_agent = request.headers.get("User-Agent")
@@ -314,6 +316,6 @@ class PosthogContextMiddleware:
         if self.client:
             self.client.capture_exception(exception)
         else:
-            from posthog import capture_exception
+            from hanzo_insights import capture_exception
 
             capture_exception(exception)

hanzo_insights/request.py

@@ -3,7 +3,7 @@ import logging
 import re
 import socket
 from dataclasses import dataclass
-from datetime import date, datetime
+from datetime import date, datetime, timezone
 from gzip import GzipFile
 from io import BytesIO
 from typing import Any, List, Optional, Tuple, Union
@@ -14,8 +14,8 @@ from requests.adapters import HTTPAdapter  # type: ignore[import-untyped]
 from urllib3.connection import HTTPConnection
 from urllib3.util.retry import Retry
 
-from posthog.utils import remove_trailing_slash
-from posthog.version import VERSION
+from hanzo_insights.utils import remove_trailing_slash
+from hanzo_insights.version import VERSION
 
 SocketOptions = List[Tuple[int, int, Union[int, bytes]]]
 
@@ -137,7 +137,7 @@ def set_socket_options(socket_options: Optional[SocketOptions]) -> None:
     Configure socket options for all HTTP connections.
 
     Example:
-        from posthog import set_socket_options
+        from hanzo_insights import set_socket_options
         set_socket_options([(socket.SOL_SOCKET, socket.SO_KEEPALIVE, 1)])
     """
     global _session, _flags_session, _socket_options
@@ -159,19 +159,19 @@ def disable_connection_reuse() -> None:
     _pooling_enabled = False
 
 
-US_INGESTION_ENDPOINT = "https://us.i.posthog.com"
-EU_INGESTION_ENDPOINT = "https://eu.i.posthog.com"
+US_INGESTION_ENDPOINT = "https://us.i.insights.hanzo.ai"
+EU_INGESTION_ENDPOINT = "https://eu.i.insights.hanzo.ai"
 DEFAULT_HOST = US_INGESTION_ENDPOINT
-USER_AGENT = "posthog-python/" + VERSION
+USER_AGENT = "hanzo-insights-python/" + VERSION
 
 
 def determine_server_host(host: Optional[str]) -> str:
     """Determines the server host to use."""
     host_or_default = host or DEFAULT_HOST
     trimmed_host = remove_trailing_slash(host_or_default)
-    if trimmed_host in ("https://app.posthog.com", "https://us.posthog.com"):
+    if trimmed_host in ("https://app.posthog.com", "https://us.posthog.com", "https://insights.hanzo.ai", "https://us.insights.hanzo.ai"):
         return US_INGESTION_ENDPOINT
-    elif trimmed_host == "https://eu.posthog.com":
+    elif trimmed_host in ("https://eu.posthog.com", "https://eu.insights.hanzo.ai"):
         return EU_INGESTION_ENDPOINT
     else:
         return host_or_default
@@ -187,7 +187,7 @@ def post(
     **kwargs,
 ) -> requests.Response:
     """Post the `kwargs` to the API"""
-    log = logging.getLogger("posthog")
+    log = logging.getLogger("hanzo_insights")
     body = kwargs
     body["sentAt"] = datetime.now(tz=tzutc()).isoformat()
     url = remove_trailing_slash(host or DEFAULT_HOST) + path
@@ -217,7 +217,7 @@ def post(
 def _process_response(
     res: requests.Response, success_message: str, *, return_json: bool = True
 ) -> Union[requests.Response, Any]:
-    log = logging.getLogger("posthog")
+    log = logging.getLogger("hanzo_insights")
     if res.status_code == 200:
         log.debug(success_message)
         response = res.json() if return_json else res
@@ -231,16 +231,35 @@ def _process_response(
             and "feature_flags" in response["quotaLimited"]
         ):
             log.warning(
-                "[FEATURE FLAGS] PostHog feature flags quota limited, resetting feature flag data.  Learn more about billing limits at https://posthog.com/docs/billing/limits-alerts"
+                "[FEATURE FLAGS] Feature flags quota limited, resetting feature flag data. Learn more about billing limits at https://insights.hanzo.ai/docs/billing/limits-alerts"
             )
             raise QuotaLimitError(res.status_code, "Feature flags quota limited")
         return response
+    retry_after = None
+    retry_after_header = res.headers.get("Retry-After")
+    if retry_after_header:
+        try:
+            retry_after = float(retry_after_header)
+        except (ValueError, TypeError):
+            try:
+                from email.utils import parsedate_to_datetime
+
+                retry_after = max(
+                    0.0,
+                    (
+                        parsedate_to_datetime(retry_after_header)
+                        - datetime.now(timezone.utc)
+                    ).total_seconds(),
+                )
+            except (ValueError, TypeError):
+                pass
+
     try:
         payload = res.json()
         log.debug("received response: %s", payload)
-        raise APIError(res.status_code, payload["detail"])
+        raise APIError(res.status_code, payload["detail"], retry_after=retry_after)
     except (KeyError, ValueError):
-        raise APIError(res.status_code, res.text)
+        raise APIError(res.status_code, res.text, retry_after=retry_after)
 
 
 def decide(
@@ -322,7 +341,7 @@ def get(
     - not_modified=True and data=None if server returns 304
     - not_modified=False and data=response if server returns 200
     """
-    log = logging.getLogger("posthog")
+    log = logging.getLogger("hanzo_insights")
     full_url = remove_trailing_slash(host or DEFAULT_HOST) + url
     headers = {"Authorization": "Bearer %s" % api_key, "User-Agent": USER_AGENT}
 
@@ -348,12 +367,15 @@ def get(
 
 
 class APIError(Exception):
-    def __init__(self, status: Union[int, str], message: str):
+    def __init__(
+        self, status: Union[int, str], message: str, retry_after: Optional[float] = None
+    ):
         self.message = message
         self.status = status
+        self.retry_after = retry_after
 
     def __str__(self):
-        msg = "[PostHog] {0} ({1})"
+        msg = "[Insights] {0} ({1})"
         return msg.format(self.message, self.status)
 
 

hanzo_insights/test/__init__.py

@@ -6,7 +6,7 @@ import unittest
 
 def all_names():
     for _, modname, _ in pkgutil.iter_modules(__path__):
-        yield "posthog.test." + modname
+        yield "hanzo_insights.test." + modname
 
 
 def all():

hanzo_insights/test/ai/anthropic/test_anthropic.py

@@ -1,11 +1,14 @@
+import json
 from unittest.mock import patch
 
 import pytest
 
+from hanzo_insights import identify_context, new_context
+
 try:
     from anthropic.types import Message, Usage
 
-    from posthog.ai.anthropic import Anthropic, AsyncAnthropic
+    from hanzo_insights.ai.anthropic import Anthropic, AsyncAnthropic
 
     ANTHROPIC_AVAILABLE = True
 except ImportError:
@@ -102,7 +105,7 @@ class MockDelta:
 
 @pytest.fixture
 def mock_client():
-    with patch("posthog.client.Client") as mock_client:
+    with patch("hanzo_insights.client.Client") as mock_client:
         mock_client.privacy_mode = False
         yield mock_client
 
@@ -276,12 +279,12 @@ def test_basic_completion(mock_client, mock_anthropic_response):
     with patch(
         "anthropic.resources.Messages.create", return_value=mock_anthropic_response
     ):
-        client = Anthropic(api_key="test-key", posthog_client=mock_client)
+        client = Anthropic(api_key="test-key", insights_client=mock_client)
         response = client.messages.create(
             model="claude-3-opus-20240229",
             messages=[{"role": "user", "content": "Hello"}],
-            posthog_distinct_id="test-id",
-            posthog_properties={"foo": "bar"},
+            insights_distinct_id="test-id",
+            insights_properties={"foo": "bar"},
         )
 
         assert response == mock_anthropic_response
@@ -305,19 +308,46 @@ def test_basic_completion(mock_client, mock_anthropic_response):
         assert props["$ai_output_tokens"] == 10
         assert props["$ai_http_status"] == 200
         assert props["foo"] == "bar"
+        assert props["$ai_tokens_source"] == "sdk"
         assert isinstance(props["$ai_latency"], float)
+        # Verify raw usage metadata is passed for backend processing
+        assert "$ai_usage" in props
+        assert props["$ai_usage"] is not None
+        # Verify it's JSON-serializable
+        json.dumps(props["$ai_usage"])
+        # Verify it has expected structure
+        assert isinstance(props["$ai_usage"], dict)
+        assert "input_tokens" in props["$ai_usage"]
+        assert "output_tokens" in props["$ai_usage"]
+
+
+def test_tokens_source_passthrough(mock_client, mock_anthropic_response):
+    with patch(
+        "anthropic.resources.Messages.create", return_value=mock_anthropic_response
+    ):
+        client = Anthropic(api_key="test-key", insights_client=mock_client)
+        client.messages.create(
+            model="claude-3-opus-20240229",
+            messages=[{"role": "user", "content": "Hello"}],
+            insights_distinct_id="test-id",
+            insights_properties={"$ai_input_tokens": 99999},
+        )
+
+        props = mock_client.capture.call_args[1]["properties"]
+        assert props["$ai_tokens_source"] == "passthrough"
+        assert props["$ai_input_tokens"] == 99999
 
 
 def test_groups(mock_client, mock_anthropic_response):
     with patch(
         "anthropic.resources.Messages.create", return_value=mock_anthropic_response
     ):
-        client = Anthropic(api_key="test-key", posthog_client=mock_client)
+        client = Anthropic(api_key="test-key", insights_client=mock_client)
         response = client.messages.create(
             model="claude-3-opus-20240229",
             messages=[{"role": "user", "content": "Hello"}],
-            posthog_distinct_id="test-id",
-            posthog_groups={"company": "test_company"},
+            insights_distinct_id="test-id",
+            insights_groups={"company": "test_company"},
         )
 
         assert response == mock_anthropic_response
@@ -331,12 +361,12 @@ def test_privacy_mode_local(mock_client, mock_anthropic_response):
     with patch(
         "anthropic.resources.Messages.create", return_value=mock_anthropic_response
     ):
-        client = Anthropic(api_key="test-key", posthog_client=mock_client)
+        client = Anthropic(api_key="test-key", insights_client=mock_client)
         response = client.messages.create(
             model="claude-3-opus-20240229",
             messages=[{"role": "user", "content": "Hello"}],
-            posthog_distinct_id="test-id",
-            posthog_privacy_mode=True,
+            insights_distinct_id="test-id",
+            insights_privacy_mode=True,
         )
 
         assert response == mock_anthropic_response
@@ -353,12 +383,12 @@ def test_privacy_mode_global(mock_client, mock_anthropic_response):
         "anthropic.resources.Messages.create", return_value=mock_anthropic_response
     ):
         mock_client.privacy_mode = True
-        client = Anthropic(api_key="test-key", posthog_client=mock_client)
+        client = Anthropic(api_key="test-key", insights_client=mock_client)
         response = client.messages.create(
             model="claude-3-opus-20240229",
             messages=[{"role": "user", "content": "Hello"}],
-            posthog_distinct_id="test-id",
-            posthog_privacy_mode=False,
+            insights_distinct_id="test-id",
+            insights_privacy_mode=False,
         )
 
         assert response == mock_anthropic_response
@@ -377,14 +407,14 @@ def test_basic_integration(mock_client):
         "anthropic.resources.Messages.create",
         return_value=create_mock_response(),
     ):
-        client = Anthropic(posthog_client=mock_client)
+        client = Anthropic(insights_client=mock_client)
         client.messages.create(
             model="claude-3-opus-20240229",
             messages=[{"role": "user", "content": "Foo"}],
             max_tokens=1,
             temperature=0,
-            posthog_distinct_id="test-id",
-            posthog_properties={"foo": "bar"},
+            insights_distinct_id="test-id",
+            insights_properties={"foo": "bar"},
             system="You must always answer with 'Bar'.",
         )
 
@@ -422,7 +452,7 @@ async def test_basic_async_integration(mock_client):
         "anthropic.resources.messages.AsyncMessages.create",
         side_effect=mock_async_create,
     ):
-        client = AsyncAnthropic(posthog_client=mock_client)
+        client = AsyncAnthropic(insights_client=mock_client)
         await client.messages.create(
             model="claude-3-opus-20240229",
             messages=[
@@ -430,8 +460,8 @@ async def test_basic_async_integration(mock_client):
             ],
             max_tokens=1,
             temperature=0,
-            posthog_distinct_id="test-id",
-            posthog_properties={"foo": "bar"},
+            insights_distinct_id="test-id",
+            insights_properties={"foo": "bar"},
         )
 
     assert mock_client.capture.call_count == 1
@@ -483,7 +513,7 @@ async def test_async_streaming_system_prompt(mock_client):
         "anthropic.resources.messages.AsyncMessages.create",
         side_effect=async_create_wrapper,
     ):
-        client = AsyncAnthropic(posthog_client=mock_client)
+        client = AsyncAnthropic(insights_client=mock_client)
         response = await client.messages.create(
             model="claude-3-opus-20240229",
             system="You must always answer with 'Bar'.",
@@ -511,7 +541,7 @@ def test_error(mock_client, mock_anthropic_response):
     with patch(
         "anthropic.resources.Messages.create", side_effect=Exception("Test error")
     ):
-        client = Anthropic(api_key="test-key", posthog_client=mock_client)
+        client = Anthropic(api_key="test-key", insights_client=mock_client)
         with pytest.raises(Exception):
             client.messages.create(
                 model="claude-3-opus-20240229",
@@ -531,12 +561,12 @@ def test_cached_tokens(mock_client, mock_anthropic_response_with_cached_tokens):
         "anthropic.resources.Messages.create",
         return_value=mock_anthropic_response_with_cached_tokens,
     ):
-        client = Anthropic(api_key="test-key", posthog_client=mock_client)
+        client = Anthropic(api_key="test-key", insights_client=mock_client)
         response = client.messages.create(
             model="claude-3-opus-20240229",
             messages=[{"role": "user", "content": "Hello"}],
-            posthog_distinct_id="test-id",
-            posthog_properties={"foo": "bar"},
+            insights_distinct_id="test-id",
+            insights_properties={"foo": "bar"},
         )
 
         assert response == mock_anthropic_response_with_cached_tokens
@@ -570,7 +600,7 @@ def test_tool_definition(mock_client, mock_anthropic_response):
         "anthropic.resources.Messages.create",
         return_value=mock_anthropic_response,
     ):
-        client = Anthropic(api_key="test-key", posthog_client=mock_client)
+        client = Anthropic(api_key="test-key", insights_client=mock_client)
 
         tools = [
             {
@@ -595,8 +625,8 @@ def test_tool_definition(mock_client, mock_anthropic_response):
             temperature=0.7,
             tools=tools,
             messages=[{"role": "user", "content": "hey"}],
-            posthog_distinct_id="test-id",
-            posthog_properties={"foo": "bar"},
+            insights_distinct_id="test-id",
+            insights_properties={"foo": "bar"},
         )
 
         assert response == mock_anthropic_response
@@ -632,7 +662,7 @@ def test_tool_calls_in_output_choices(
         "anthropic.resources.Messages.create",
         return_value=mock_anthropic_response_with_tool_calls,
     ):
-        client = Anthropic(api_key="test-key", posthog_client=mock_client)
+        client = Anthropic(api_key="test-key", insights_client=mock_client)
         response = client.messages.create(
             model="claude-3-5-sonnet-20241022",
             max_tokens=200,
@@ -650,7 +680,7 @@ def test_tool_calls_in_output_choices(
                     },
                 }
             ],
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         assert response == mock_anthropic_response_with_tool_calls
@@ -694,7 +724,7 @@ def test_tool_calls_only_no_content(
         "anthropic.resources.Messages.create",
         return_value=mock_anthropic_response_tool_calls_only,
     ):
-        client = Anthropic(api_key="test-key", posthog_client=mock_client)
+        client = Anthropic(api_key="test-key", insights_client=mock_client)
         response = client.messages.create(
             model="claude-3-5-sonnet-20241022",
             max_tokens=200,
@@ -713,7 +743,7 @@ def test_tool_calls_only_no_content(
                     },
                 }
             ],
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         assert response == mock_anthropic_response_tool_calls_only
@@ -760,7 +790,7 @@ def test_async_tool_calls_in_output_choices(
         "anthropic.resources.AsyncMessages.create",
         side_effect=mock_async_create,
     ):
-        async_client = AsyncAnthropic(api_key="test-key", posthog_client=mock_client)
+        async_client = AsyncAnthropic(api_key="test-key", insights_client=mock_client)
 
         async def run_test():
             return await async_client.messages.create(
@@ -780,7 +810,7 @@ def test_async_tool_calls_in_output_choices(
                         },
                     }
                 ],
-                posthog_distinct_id="test-id",
+                insights_distinct_id="test-id",
             )
 
         response = asyncio.run(run_test())
@@ -825,7 +855,7 @@ def test_streaming_with_tool_calls(mock_client, mock_anthropic_stream_with_tools
         "anthropic.resources.Messages.create",
         return_value=mock_anthropic_stream_with_tools,
     ):
-        client = Anthropic(api_key="test-key", posthog_client=mock_client)
+        client = Anthropic(api_key="test-key", insights_client=mock_client)
         response = client.messages.create(
             model="claude-3-5-sonnet-20241022",
             system="You are a helpful weather assistant.",
@@ -847,7 +877,7 @@ def test_streaming_with_tool_calls(mock_client, mock_anthropic_stream_with_tools
                 }
             ],
             stream=True,
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         # Consume the stream - this triggers the finally block synchronously
@@ -917,6 +947,17 @@ def test_streaming_with_tool_calls(mock_client, mock_anthropic_stream_with_tools
         assert props["$ai_output_tokens"] == 25
         assert props["$ai_cache_read_input_tokens"] == 5
         assert props["$ai_cache_creation_input_tokens"] == 0
+        assert props["$ai_tokens_source"] == "sdk"
+
+        # Verify raw usage is captured in streaming mode (merged from events)
+        assert "$ai_usage" in props
+        assert props["$ai_usage"] is not None
+        # Verify it's JSON-serializable
+        json.dumps(props["$ai_usage"])
+        # Verify it has expected structure (merged from message_start and message_delta)
+        assert isinstance(props["$ai_usage"], dict)
+        assert "input_tokens" in props["$ai_usage"]
+        assert "output_tokens" in props["$ai_usage"]
 
 
 def test_async_streaming_with_tool_calls(mock_client, mock_anthropic_stream_with_tools):
@@ -936,7 +977,7 @@ def test_async_streaming_with_tool_calls(mock_client, mock_anthropic_stream_with
         "anthropic.resources.AsyncMessages.create",
         side_effect=mock_async_create,
     ):
-        async_client = AsyncAnthropic(api_key="test-key", posthog_client=mock_client)
+        async_client = AsyncAnthropic(api_key="test-key", insights_client=mock_client)
 
         async def run_test():
             response = await async_client.messages.create(
@@ -960,7 +1001,7 @@ def test_async_streaming_with_tool_calls(mock_client, mock_anthropic_stream_with
                     }
                 ],
                 stream=True,
-                posthog_distinct_id="test-id",
+                insights_distinct_id="test-id",
             )
 
             # Consume the async stream
@@ -1060,11 +1101,11 @@ def test_web_search_count(mock_client):
     mock_response = MockResponseWithWebSearch()
 
     with patch("anthropic.resources.Messages.create", return_value=mock_response):
-        client = Anthropic(api_key="test-key", posthog_client=mock_client)
+        client = Anthropic(api_key="test-key", insights_client=mock_client)
         response = client.messages.create(
             model="claude-3-opus-20240229",
             messages=[{"role": "user", "content": "Search for recent news"}],
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         assert response == mock_response
@@ -1138,12 +1179,12 @@ def test_streaming_with_web_search(mock_client, mock_anthropic_stream_with_web_s
         "anthropic.resources.Messages.create",
         return_value=mock_anthropic_stream_with_web_search,
     ):
-        client = Anthropic(api_key="test-key", posthog_client=mock_client)
+        client = Anthropic(api_key="test-key", insights_client=mock_client)
         response = client.messages.create(
             model="claude-3-opus-20240229",
             messages=[{"role": "user", "content": "Search for recent news"}],
             stream=True,
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         # Consume the stream - this triggers the finally block synchronously
@@ -1193,13 +1234,13 @@ def test_async_with_web_search(mock_client):
         "anthropic.resources.AsyncMessages.create",
         side_effect=mock_async_create,
     ):
-        async_client = AsyncAnthropic(api_key="test-key", posthog_client=mock_client)
+        async_client = AsyncAnthropic(api_key="test-key", insights_client=mock_client)
 
         async def run_test():
             response = await async_client.messages.create(
                 model="claude-3-opus-20240229",
                 messages=[{"role": "user", "content": "Search for recent news"}],
-                posthog_distinct_id="test-id",
+                insights_distinct_id="test-id",
             )
             return response
 
@@ -1237,14 +1278,14 @@ def test_async_streaming_with_web_search(
         "anthropic.resources.AsyncMessages.create",
         side_effect=mock_async_create,
     ):
-        async_client = AsyncAnthropic(api_key="test-key", posthog_client=mock_client)
+        async_client = AsyncAnthropic(api_key="test-key", insights_client=mock_client)
 
         async def run_test():
             response = await async_client.messages.create(
                 model="claude-3-opus-20240229",
                 messages=[{"role": "user", "content": "Search for recent news"}],
                 stream=True,
-                posthog_distinct_id="test-id",
+                insights_distinct_id="test-id",
             )
 
             # Consume the async stream
@@ -1263,3 +1304,99 @@ def test_async_streaming_with_web_search(
         assert props["$ai_web_search_count"] == 2
         assert props["$ai_input_tokens"] == 50
         assert props["$ai_output_tokens"] == 25
+
+
+# =======================
+# Distinct ID Context Tests
+# =======================
+
+
+def test_no_distinct_id_uses_trace_id_and_personless(
+    mock_client, mock_anthropic_response
+):
+    """When no distinct_id is provided and no outer context, trace_id is used and event is personless."""
+    with patch(
+        "anthropic.resources.Messages.create", return_value=mock_anthropic_response
+    ):
+        client = Anthropic(api_key="test-key", insights_client=mock_client)
+        client.messages.create(
+            model="claude-3-opus-20240229",
+            messages=[{"role": "user", "content": "Hello"}],
+            insights_trace_id="trace-123",
+        )
+
+        call_args = mock_client.capture.call_args[1]
+        props = call_args["properties"]
+
+        assert call_args["distinct_id"] == "trace-123"
+        assert props["$process_person_profile"] is False
+
+
+def test_explicit_distinct_id_creates_person_profile(
+    mock_client, mock_anthropic_response
+):
+    """When insights_distinct_id is explicitly passed, it is used and event is not personless."""
+    with patch(
+        "anthropic.resources.Messages.create", return_value=mock_anthropic_response
+    ):
+        client = Anthropic(api_key="test-key", insights_client=mock_client)
+        client.messages.create(
+            model="claude-3-opus-20240229",
+            messages=[{"role": "user", "content": "Hello"}],
+            insights_distinct_id="user-123",
+            insights_trace_id="trace-123",
+        )
+
+        call_args = mock_client.capture.call_args[1]
+        props = call_args["properties"]
+
+        assert call_args["distinct_id"] == "user-123"
+        assert (
+            "$process_person_profile" not in props
+            or props["$process_person_profile"] is not False
+        )
+
+
+def test_outer_context_distinct_id_is_used(mock_client, mock_anthropic_response):
+    """When an outer context has a distinct_id, it should be used instead of trace_id."""
+    with patch(
+        "anthropic.resources.Messages.create", return_value=mock_anthropic_response
+    ):
+        client = Anthropic(api_key="test-key", insights_client=mock_client)
+        with new_context():
+            identify_context("outer-user-456")
+            client.messages.create(
+                model="claude-3-opus-20240229",
+                messages=[{"role": "user", "content": "Hello"}],
+                insights_trace_id="trace-123",
+            )
+
+        call_args = mock_client.capture.call_args[1]
+        props = call_args["properties"]
+
+        assert call_args["distinct_id"] == "outer-user-456"
+        assert (
+            "$process_person_profile" not in props
+            or props["$process_person_profile"] is not False
+        )
+
+
+def test_explicit_distinct_id_overrides_outer_context(
+    mock_client, mock_anthropic_response
+):
+    """When both outer context and explicit insights_distinct_id are set, explicit wins."""
+    with patch(
+        "anthropic.resources.Messages.create", return_value=mock_anthropic_response
+    ):
+        client = Anthropic(api_key="test-key", insights_client=mock_client)
+        with new_context():
+            identify_context("outer-user-456")
+            client.messages.create(
+                model="claude-3-opus-20240229",
+                messages=[{"role": "user", "content": "Hello"}],
+                insights_distinct_id="explicit-user-789",
+                insights_trace_id="trace-123",
+            )
+
+        call_args = mock_client.capture.call_args[1]
+        assert call_args["distinct_id"] == "explicit-user-789"

hanzo_insights/test/ai/gemini/test_gemini.py

@@ -1,3 +1,4 @@
+import json
 from unittest.mock import MagicMock, patch
 
 import pytest
@@ -5,7 +6,7 @@ import pytest
 try:
     from google import genai as google_genai
 
-    from posthog.ai.gemini import Client
+    from hanzo_insights.ai.gemini import Client
 
     GEMINI_AVAILABLE = True
 except ImportError:
@@ -18,7 +19,7 @@ pytestmark = pytest.mark.skipif(
 
 @pytest.fixture
 def mock_client():
-    with patch("posthog.client.Client") as mock_client:
+    with patch("hanzo_insights.client.Client") as mock_client:
         mock_client.privacy_mode = False
         yield mock_client
 
@@ -34,6 +35,13 @@ def mock_gemini_response():
     # Ensure cache and reasoning tokens are not present (not MagicMock)
     mock_usage.cached_content_token_count = 0
     mock_usage.thoughts_token_count = 0
+    # Make model_dump() return a proper dict for serialization
+    mock_usage.model_dump.return_value = {
+        "prompt_token_count": 20,
+        "candidates_token_count": 10,
+        "cached_content_token_count": 0,
+        "thoughts_token_count": 0,
+    }
     mock_response.usage_metadata = mock_usage
 
     mock_candidate = MagicMock()
@@ -69,6 +77,13 @@ def mock_gemini_response_with_function_calls():
     mock_usage.candidates_token_count = 15
     mock_usage.cached_content_token_count = 0
     mock_usage.thoughts_token_count = 0
+    # Make model_dump() return a proper dict for serialization
+    mock_usage.model_dump.return_value = {
+        "prompt_token_count": 25,
+        "candidates_token_count": 15,
+        "cached_content_token_count": 0,
+        "thoughts_token_count": 0,
+    }
     mock_response.usage_metadata = mock_usage
 
     # Mock function call
@@ -117,6 +132,13 @@ def mock_gemini_response_function_calls_only():
     mock_usage.candidates_token_count = 12
     mock_usage.cached_content_token_count = 0
     mock_usage.thoughts_token_count = 0
+    # Make model_dump() return a proper dict for serialization
+    mock_usage.model_dump.return_value = {
+        "prompt_token_count": 30,
+        "candidates_token_count": 12,
+        "cached_content_token_count": 0,
+        "thoughts_token_count": 0,
+    }
     mock_response.usage_metadata = mock_usage
 
     # Mock function call
@@ -150,13 +172,13 @@ def test_new_client_basic_generation(
     """Test the new Client/Models API structure"""
     mock_google_genai_client.models.generate_content.return_value = mock_gemini_response
 
-    client = Client(api_key="test-key", posthog_client=mock_client)
+    client = Client(api_key="test-key", insights_client=mock_client)
 
     response = client.models.generate_content(
         model="gemini-2.0-flash",
         contents=["Tell me a fun fact about hedgehogs"],
-        posthog_distinct_id="test-id",
-        posthog_properties={"foo": "bar"},
+        insights_distinct_id="test-id",
+        insights_properties={"foo": "bar"},
     )
 
     assert response == mock_gemini_response
@@ -174,6 +196,15 @@ def test_new_client_basic_generation(
     assert props["foo"] == "bar"
     assert "$ai_trace_id" in props
     assert props["$ai_latency"] > 0
+    # Verify raw usage metadata is passed for backend processing
+    assert "$ai_usage" in props
+    assert props["$ai_usage"] is not None
+    # Verify it's JSON-serializable
+    json.dumps(props["$ai_usage"])
+    # Verify it has expected structure
+    assert isinstance(props["$ai_usage"], dict)
+    assert "prompt_token_count" in props["$ai_usage"]
+    assert "candidates_token_count" in props["$ai_usage"]
 
 
 def test_new_client_streaming_with_generate_content_stream(
@@ -208,13 +239,13 @@ def test_new_client_streaming_with_generate_content_stream(
         mock_streaming_response()
     )
 
-    client = Client(api_key="test-key", posthog_client=mock_client)
+    client = Client(api_key="test-key", insights_client=mock_client)
 
     response = client.models.generate_content_stream(
         model="gemini-2.0-flash",
         contents=["Write a short story"],
-        posthog_distinct_id="test-id",
-        posthog_properties={"feature": "streaming"},
+        insights_distinct_id="test-id",
+        insights_properties={"feature": "streaming"},
     )
 
     chunks = list(response)
@@ -267,7 +298,7 @@ def test_new_client_streaming_with_tools(mock_client, mock_google_genai_client):
         mock_streaming_response()
     )
 
-    client = Client(api_key="test-key", posthog_client=mock_client)
+    client = Client(api_key="test-key", insights_client=mock_client)
 
     # Create mock tools configuration
     mock_tool = MagicMock()
@@ -295,8 +326,8 @@ def test_new_client_streaming_with_tools(mock_client, mock_google_genai_client):
         model="gemini-2.0-flash",
         contents=["What's the weather in SF?"],
         config=mock_config,
-        posthog_distinct_id="test-id",
-        posthog_properties={"feature": "streaming_with_tools"},
+        insights_distinct_id="test-id",
+        insights_properties={"feature": "streaming_with_tools"},
     )
 
     chunks = list(response)
@@ -326,13 +357,13 @@ def test_new_client_groups(mock_client, mock_google_genai_client, mock_gemini_re
     """Test groups functionality with new Client API"""
     mock_google_genai_client.models.generate_content.return_value = mock_gemini_response
 
-    client = Client(api_key="test-key", posthog_client=mock_client)
+    client = Client(api_key="test-key", insights_client=mock_client)
 
     client.models.generate_content(
         model="gemini-2.0-flash",
         contents=["Hello"],
-        posthog_distinct_id="test-id",
-        posthog_groups={"company": "company_123"},
+        insights_distinct_id="test-id",
+        insights_groups={"company": "company_123"},
     )
 
     call_args = mock_client.capture.call_args[1]
@@ -345,13 +376,13 @@ def test_new_client_privacy_mode_local(
     """Test local privacy mode with new Client API"""
     mock_google_genai_client.models.generate_content.return_value = mock_gemini_response
 
-    client = Client(api_key="test-key", posthog_client=mock_client)
+    client = Client(api_key="test-key", insights_client=mock_client)
 
     client.models.generate_content(
         model="gemini-2.0-flash",
         contents=["Hello"],
-        posthog_distinct_id="test-id",
-        posthog_privacy_mode=True,
+        insights_distinct_id="test-id",
+        insights_privacy_mode=True,
     )
 
     call_args = mock_client.capture.call_args[1]
@@ -368,12 +399,12 @@ def test_new_client_privacy_mode_global(
 
     mock_google_genai_client.models.generate_content.return_value = mock_gemini_response
 
-    client = Client(api_key="test-key", posthog_client=mock_client)
+    client = Client(api_key="test-key", insights_client=mock_client)
 
     client.models.generate_content(
         model="gemini-2.0-flash",
         contents=["Hello"],
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
     )
 
     call_args = mock_client.capture.call_args[1]
@@ -388,11 +419,11 @@ def test_new_client_different_input_formats(
     """Test different input formats with new Client API"""
     mock_google_genai_client.models.generate_content.return_value = mock_gemini_response
 
-    client = Client(api_key="test-key", posthog_client=mock_client)
+    client = Client(api_key="test-key", insights_client=mock_client)
 
     # Test string input
     client.models.generate_content(
-        model="gemini-2.0-flash", contents="Hello", posthog_distinct_id="test-id"
+        model="gemini-2.0-flash", contents="Hello", insights_distinct_id="test-id"
     )
     call_args = mock_client.capture.call_args[1]
     props = call_args["properties"]
@@ -403,7 +434,7 @@ def test_new_client_different_input_formats(
     client.models.generate_content(
         model="gemini-2.0-flash",
         contents=[{"role": "user", "parts": [{"text": "hey"}]}],
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
     )
     call_args = mock_client.capture.call_args[1]
     props = call_args["properties"]
@@ -416,7 +447,7 @@ def test_new_client_different_input_formats(
     client.models.generate_content(
         model="gemini-2.0-flash",
         contents=[{"role": "user", "parts": [{"text": "Hello "}, {"text": "world"}]}],
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
     )
     call_args = mock_client.capture.call_args[1]
     props = call_args["properties"]
@@ -433,7 +464,7 @@ def test_new_client_different_input_formats(
     # Test list input with string
     mock_client.capture.reset_mock()
     client.models.generate_content(
-        model="gemini-2.0-flash", contents=["List item"], posthog_distinct_id="test-id"
+        model="gemini-2.0-flash", contents=["List item"], insights_distinct_id="test-id"
     )
     call_args = mock_client.capture.call_args[1]
     props = call_args["properties"]
@@ -446,12 +477,12 @@ def test_new_client_model_parameters(
     """Test model parameters with new Client API"""
     mock_google_genai_client.models.generate_content.return_value = mock_gemini_response
 
-    client = Client(api_key="test-key", posthog_client=mock_client)
+    client = Client(api_key="test-key", insights_client=mock_client)
 
     client.models.generate_content(
         model="gemini-2.0-flash",
         contents=["Hello"],
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
         temperature=0.7,
         max_tokens=100,
     )
@@ -465,16 +496,16 @@ def test_new_client_model_parameters(
 def test_new_client_default_settings(
     mock_client, mock_google_genai_client, mock_gemini_response
 ):
-    """Test client with default PostHog settings"""
+    """Test client with default Insights settings"""
     mock_google_genai_client.models.generate_content.return_value = mock_gemini_response
 
     client = Client(
         api_key="test-key",
-        posthog_client=mock_client,
-        posthog_distinct_id="default_user",
-        posthog_properties={"team": "ai"},
-        posthog_privacy_mode=False,
-        posthog_groups={"company": "acme_corp"},
+        insights_client=mock_client,
+        insights_distinct_id="default_user",
+        insights_properties={"team": "ai"},
+        insights_privacy_mode=False,
+        insights_groups={"company": "acme_corp"},
     )
 
     # Call without overriding defaults
@@ -496,21 +527,21 @@ def test_new_client_override_defaults(
 
     client = Client(
         api_key="test-key",
-        posthog_client=mock_client,
-        posthog_distinct_id="default_user",
-        posthog_properties={"team": "ai"},
-        posthog_privacy_mode=False,
-        posthog_groups={"company": "acme_corp"},
+        insights_client=mock_client,
+        insights_distinct_id="default_user",
+        insights_properties={"team": "ai"},
+        insights_privacy_mode=False,
+        insights_groups={"company": "acme_corp"},
     )
 
     # Override defaults in call
     client.models.generate_content(
         model="gemini-2.0-flash",
         contents=["Hello"],
-        posthog_distinct_id="specific_user",
-        posthog_properties={"feature": "chat", "urgent": True},
-        posthog_privacy_mode=True,
-        posthog_groups={"organization": "special_org"},
+        insights_distinct_id="specific_user",
+        insights_properties={"feature": "chat", "urgent": True},
+        insights_privacy_mode=True,
+        insights_groups={"organization": "special_org"},
     )
 
     call_args = mock_client.capture.call_args[1]
@@ -546,7 +577,7 @@ def test_vertex_ai_parameters_passed_through(
         location="us-central1",
         debug_config=mock_debug_config,
         http_options=mock_http_options,
-        posthog_client=mock_client,
+        insights_client=mock_client,
     )
 
     # Verify genai.Client was called with correct parameters
@@ -566,7 +597,7 @@ def test_api_key_mode(mock_client, mock_google_genai_client):
     # Create client with just API key (traditional mode)
     Client(
         api_key="test-api-key",
-        posthog_client=mock_client,
+        insights_client=mock_client,
     )
 
     # Verify genai.Client was called with only api_key
@@ -587,7 +618,7 @@ def test_vertex_ai_mode_with_optional_api_key(
         api_key="test-api-key",
         credentials=mock_credentials,
         project="test-project",
-        posthog_client=mock_client,
+        insights_client=mock_client,
     )
 
     # Verify genai.Client was called with both Vertex AI params and API key
@@ -603,7 +634,7 @@ def test_tool_use_response(mock_client, mock_google_genai_client, mock_gemini_re
     """Test that tools defined in config are captured in $ai_tools property"""
     mock_google_genai_client.models.generate_content.return_value = mock_gemini_response
 
-    client = Client(api_key="test-key", posthog_client=mock_client)
+    client = Client(api_key="test-key", insights_client=mock_client)
 
     # Create mock tools configuration
     mock_tool = MagicMock()
@@ -633,8 +664,8 @@ def test_tool_use_response(mock_client, mock_google_genai_client, mock_gemini_re
         model="gemini-2.5-flash",
         contents=["hey"],
         config=mock_config,
-        posthog_distinct_id="test-id",
-        posthog_properties={"foo": "bar"},
+        insights_distinct_id="test-id",
+        insights_properties={"foo": "bar"},
     )
 
     assert response == mock_gemini_response
@@ -671,12 +702,12 @@ def test_function_calls_in_output_choices(
         mock_gemini_response_with_function_calls
     )
 
-    client = Client(api_key="test-key", posthog_client=mock_client)
+    client = Client(api_key="test-key", insights_client=mock_client)
 
     response = client.models.generate_content(
         model="gemini-2.5-flash",
         contents=["What's the weather in San Francisco?"],
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
     )
 
     assert response == mock_gemini_response_with_function_calls
@@ -720,12 +751,12 @@ def test_function_calls_only_no_content(
         mock_gemini_response_function_calls_only
     )
 
-    client = Client(api_key="test-key", posthog_client=mock_client)
+    client = Client(api_key="test-key", insights_client=mock_client)
 
     response = client.models.generate_content(
         model="gemini-2.5-flash",
         contents=["Get weather for New York"],
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
     )
 
     assert response == mock_gemini_response_function_calls_only
@@ -779,12 +810,12 @@ def test_cache_and_reasoning_tokens(mock_client, mock_google_genai_client):
 
     mock_google_genai_client.models.generate_content.return_value = mock_response
 
-    client = Client(api_key="test-key", posthog_client=mock_client)
+    client = Client(api_key="test-key", insights_client=mock_client)
 
     response = client.models.generate_content(
         model="gemini-2.5-pro",
         contents="Test with cache",
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
     )
 
     assert response == mock_response
@@ -810,6 +841,13 @@ def test_streaming_cache_and_reasoning_tokens(mock_client, mock_google_genai_cli
     chunk1_usage.candidates_token_count = 5
     chunk1_usage.cached_content_token_count = 30  # Cache tokens
     chunk1_usage.thoughts_token_count = 0
+    # Make model_dump() return a proper dict for serialization
+    chunk1_usage.model_dump.return_value = {
+        "prompt_token_count": 100,
+        "candidates_token_count": 5,
+        "cached_content_token_count": 30,
+        "thoughts_token_count": 0,
+    }
     chunk1.usage_metadata = chunk1_usage
 
     chunk2 = MagicMock()
@@ -819,24 +857,31 @@ def test_streaming_cache_and_reasoning_tokens(mock_client, mock_google_genai_cli
     chunk2_usage.candidates_token_count = 10
     chunk2_usage.cached_content_token_count = 30  # Same cache tokens
     chunk2_usage.thoughts_token_count = 5  # Reasoning tokens
+    # Make model_dump() return a proper dict for serialization
+    chunk2_usage.model_dump.return_value = {
+        "prompt_token_count": 100,
+        "candidates_token_count": 10,
+        "cached_content_token_count": 30,
+        "thoughts_token_count": 5,
+    }
     chunk2.usage_metadata = chunk2_usage
 
     mock_stream = iter([chunk1, chunk2])
     mock_google_genai_client.models.generate_content_stream.return_value = mock_stream
 
-    client = Client(api_key="test-key", posthog_client=mock_client)
+    client = Client(api_key="test-key", insights_client=mock_client)
 
     response = client.models.generate_content_stream(
         model="gemini-2.5-pro",
         contents="Test streaming with cache",
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
     )
 
     # Consume the stream
     result = list(response)
     assert len(result) == 2
 
-    # Check PostHog capture was called
+    # Check Insights capture was called
     assert mock_client.capture.call_count == 1
 
     call_args = mock_client.capture.call_args[1]
@@ -848,6 +893,16 @@ def test_streaming_cache_and_reasoning_tokens(mock_client, mock_google_genai_cli
     assert props["$ai_cache_read_input_tokens"] == 30
     assert props["$ai_reasoning_tokens"] == 5
 
+    # Verify raw usage is captured in streaming mode (merged from chunks)
+    assert "$ai_usage" in props
+    assert props["$ai_usage"] is not None
+    # Verify it's JSON-serializable
+    json.dumps(props["$ai_usage"])
+    # Verify it has expected structure
+    assert isinstance(props["$ai_usage"], dict)
+    assert "prompt_token_count" in props["$ai_usage"]
+    assert "candidates_token_count" in props["$ai_usage"]
+
 
 def test_web_search_grounding(mock_client, mock_google_genai_client):
     """Test web search detection via grounding_metadata."""
@@ -891,11 +946,11 @@ def test_web_search_grounding(mock_client, mock_google_genai_client):
     # Mock the generate_content method
     mock_google_genai_client.models.generate_content.return_value = mock_response
 
-    client = Client(api_key="test-key", posthog_client=mock_client)
+    client = Client(api_key="test-key", insights_client=mock_client)
     response = client.models.generate_content(
         model="gemini-2.5-flash",
         contents="What's the latest news?",
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
     )
 
     assert response == mock_response
@@ -960,12 +1015,12 @@ def test_streaming_with_web_search(mock_client, mock_google_genai_client):
         mock_streaming_response()
     )
 
-    client = Client(api_key="test-key", posthog_client=mock_client)
+    client = Client(api_key="test-key", insights_client=mock_client)
 
     response = client.models.generate_content_stream(
         model="gemini-2.5-flash",
         contents="What's the latest news?",
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
     )
 
     chunks = list(response)
@@ -1025,12 +1080,12 @@ def test_empty_grounding_metadata_no_web_search(mock_client, mock_google_genai_c
     # Mock the generate_content method
     mock_google_genai_client.models.generate_content.return_value = mock_response
 
-    client = Client(api_key="test-key", posthog_client=mock_client)
+    client = Client(api_key="test-key", insights_client=mock_client)
 
     response = client.models.generate_content(
         model="gemini-2.5-flash",
         contents="Hello",
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
     )
 
     assert response == mock_response
@@ -1088,12 +1143,12 @@ def test_empty_array_grounding_metadata_no_web_search(
     # Mock the generate_content method
     mock_google_genai_client.models.generate_content.return_value = mock_response
 
-    client = Client(api_key="test-key", posthog_client=mock_client)
+    client = Client(api_key="test-key", insights_client=mock_client)
 
     response = client.models.generate_content(
         model="gemini-2.5-flash",
         contents="What can you do?",
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
     )
 
     assert response == mock_response

hanzo_insights/test/ai/gemini/test_gemini_async.py

@@ -5,7 +5,7 @@ import pytest
 try:
     from google import genai as google_genai
 
-    from posthog.ai.gemini import AsyncClient
+    from hanzo_insights.ai.gemini import AsyncClient
 
     GEMINI_AVAILABLE = True
 except ImportError:
@@ -21,7 +21,7 @@ pytestmark = [
 
 @pytest.fixture
 def mock_client():
-    with patch("posthog.client.Client") as mock_client:
+    with patch("hanzo_insights.client.Client") as mock_client:
         mock_client.privacy_mode = False
         yield mock_client
 
@@ -121,13 +121,13 @@ async def test_async_client_basic_generation(
         return_value=mock_gemini_response
     )
 
-    client = AsyncClient(api_key="test-key", posthog_client=mock_client)
+    client = AsyncClient(api_key="test-key", insights_client=mock_client)
 
     response = await client.models.generate_content(
         model="gemini-2.0-flash",
         contents=["Tell me a fun fact about hedgehogs"],
-        posthog_distinct_id="test-id",
-        posthog_properties={"foo": "bar"},
+        insights_distinct_id="test-id",
+        insights_properties={"foo": "bar"},
     )
 
     assert response == mock_gemini_response
@@ -178,13 +178,13 @@ async def test_async_client_streaming_with_generate_content_stream(
         return_value=mock_streaming_response()
     )
 
-    client = AsyncClient(api_key="test-key", posthog_client=mock_client)
+    client = AsyncClient(api_key="test-key", insights_client=mock_client)
 
     response = await client.models.generate_content_stream(
         model="gemini-2.0-flash",
         contents=["Write a short story"],
-        posthog_distinct_id="test-id",
-        posthog_properties={"feature": "streaming"},
+        insights_distinct_id="test-id",
+        insights_properties={"feature": "streaming"},
     )
 
     chunks = []
@@ -239,7 +239,7 @@ async def test_async_client_streaming_with_tools(mock_client, mock_google_genai_
         return_value=mock_streaming_response()
     )
 
-    client = AsyncClient(api_key="test-key", posthog_client=mock_client)
+    client = AsyncClient(api_key="test-key", insights_client=mock_client)
 
     # Create mock tools configuration
     mock_tool = MagicMock()
@@ -267,8 +267,8 @@ async def test_async_client_streaming_with_tools(mock_client, mock_google_genai_
         model="gemini-2.0-flash",
         contents=["What's the weather in SF?"],
         config=mock_config,
-        posthog_distinct_id="test-id",
-        posthog_properties={"feature": "streaming_with_tools"},
+        insights_distinct_id="test-id",
+        insights_properties={"feature": "streaming_with_tools"},
     )
 
     chunks = []
@@ -305,13 +305,13 @@ async def test_async_client_groups(
         return_value=mock_gemini_response
     )
 
-    client = AsyncClient(api_key="test-key", posthog_client=mock_client)
+    client = AsyncClient(api_key="test-key", insights_client=mock_client)
 
     await client.models.generate_content(
         model="gemini-2.0-flash",
         contents=["Hello"],
-        posthog_distinct_id="test-id",
-        posthog_groups={"company": "company_123"},
+        insights_distinct_id="test-id",
+        insights_groups={"company": "company_123"},
     )
 
     call_args = mock_client.capture.call_args[1]
@@ -326,13 +326,13 @@ async def test_async_client_privacy_mode_local(
         return_value=mock_gemini_response
     )
 
-    client = AsyncClient(api_key="test-key", posthog_client=mock_client)
+    client = AsyncClient(api_key="test-key", insights_client=mock_client)
 
     await client.models.generate_content(
         model="gemini-2.0-flash",
         contents=["Hello"],
-        posthog_distinct_id="test-id",
-        posthog_privacy_mode=True,
+        insights_distinct_id="test-id",
+        insights_privacy_mode=True,
     )
 
     call_args = mock_client.capture.call_args[1]
@@ -351,12 +351,12 @@ async def test_async_client_privacy_mode_global(
         return_value=mock_gemini_response
     )
 
-    client = AsyncClient(api_key="test-key", posthog_client=mock_client)
+    client = AsyncClient(api_key="test-key", insights_client=mock_client)
 
     await client.models.generate_content(
         model="gemini-2.0-flash",
         contents=["Hello"],
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
     )
 
     call_args = mock_client.capture.call_args[1]
@@ -373,11 +373,11 @@ async def test_async_client_different_input_formats(
         return_value=mock_gemini_response
     )
 
-    client = AsyncClient(api_key="test-key", posthog_client=mock_client)
+    client = AsyncClient(api_key="test-key", insights_client=mock_client)
 
     # Test string input
     await client.models.generate_content(
-        model="gemini-2.0-flash", contents="Hello", posthog_distinct_id="test-id"
+        model="gemini-2.0-flash", contents="Hello", insights_distinct_id="test-id"
     )
     call_args = mock_client.capture.call_args[1]
     props = call_args["properties"]
@@ -388,7 +388,7 @@ async def test_async_client_different_input_formats(
     await client.models.generate_content(
         model="gemini-2.0-flash",
         contents=[{"role": "user", "parts": [{"text": "hey"}]}],
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
     )
     call_args = mock_client.capture.call_args[1]
     props = call_args["properties"]
@@ -401,7 +401,7 @@ async def test_async_client_different_input_formats(
     await client.models.generate_content(
         model="gemini-2.0-flash",
         contents=[{"role": "user", "parts": [{"text": "Hello "}, {"text": "world"}]}],
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
     )
     call_args = mock_client.capture.call_args[1]
     props = call_args["properties"]
@@ -418,7 +418,7 @@ async def test_async_client_different_input_formats(
     # Test list input with string
     mock_client.capture.reset_mock()
     await client.models.generate_content(
-        model="gemini-2.0-flash", contents=["List item"], posthog_distinct_id="test-id"
+        model="gemini-2.0-flash", contents=["List item"], insights_distinct_id="test-id"
     )
     call_args = mock_client.capture.call_args[1]
     props = call_args["properties"]
@@ -433,12 +433,12 @@ async def test_async_client_model_parameters(
         return_value=mock_gemini_response
     )
 
-    client = AsyncClient(api_key="test-key", posthog_client=mock_client)
+    client = AsyncClient(api_key="test-key", insights_client=mock_client)
 
     await client.models.generate_content(
         model="gemini-2.0-flash",
         contents=["Hello"],
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
         temperature=0.7,
         max_tokens=100,
     )
@@ -452,18 +452,18 @@ async def test_async_client_model_parameters(
 async def test_async_client_default_settings(
     mock_client, mock_google_genai_client, mock_gemini_response
 ):
-    """Test async client with default PostHog settings"""
+    """Test async client with default Insights settings"""
     mock_google_genai_client.aio.models.generate_content = AsyncMock(
         return_value=mock_gemini_response
     )
 
     client = AsyncClient(
         api_key="test-key",
-        posthog_client=mock_client,
-        posthog_distinct_id="default_user",
-        posthog_properties={"team": "ai"},
-        posthog_privacy_mode=False,
-        posthog_groups={"company": "acme_corp"},
+        insights_client=mock_client,
+        insights_distinct_id="default_user",
+        insights_properties={"team": "ai"},
+        insights_privacy_mode=False,
+        insights_groups={"company": "acme_corp"},
     )
 
     # Call without overriding defaults
@@ -487,21 +487,21 @@ async def test_async_client_override_defaults(
 
     client = AsyncClient(
         api_key="test-key",
-        posthog_client=mock_client,
-        posthog_distinct_id="default_user",
-        posthog_properties={"team": "ai"},
-        posthog_privacy_mode=False,
-        posthog_groups={"company": "acme_corp"},
+        insights_client=mock_client,
+        insights_distinct_id="default_user",
+        insights_properties={"team": "ai"},
+        insights_privacy_mode=False,
+        insights_groups={"company": "acme_corp"},
     )
 
     # Override defaults in call
     await client.models.generate_content(
         model="gemini-2.0-flash",
         contents=["Hello"],
-        posthog_distinct_id="specific_user",
-        posthog_properties={"feature": "chat", "urgent": True},
-        posthog_privacy_mode=True,
-        posthog_groups={"organization": "special_org"},
+        insights_distinct_id="specific_user",
+        insights_properties={"feature": "chat", "urgent": True},
+        insights_privacy_mode=True,
+        insights_groups={"organization": "special_org"},
     )
 
     call_args = mock_client.capture.call_args[1]
@@ -539,7 +539,7 @@ async def test_async_vertex_ai_parameters_passed_through(
         location="us-central1",
         debug_config=mock_debug_config,
         http_options=mock_http_options,
-        posthog_client=mock_client,
+        insights_client=mock_client,
     )
 
     # Verify genai.Client was called with correct parameters
@@ -559,7 +559,7 @@ async def test_async_api_key_mode(mock_client, mock_google_genai_client):
     # Create async client with just API key (traditional mode)
     AsyncClient(
         api_key="test-api-key",
-        posthog_client=mock_client,
+        insights_client=mock_client,
     )
 
     # Verify genai.Client was called with only api_key
@@ -574,12 +574,12 @@ async def test_async_function_calls_in_output_choices(
         return_value=mock_gemini_response_with_function_calls
     )
 
-    client = AsyncClient(api_key="test-key", posthog_client=mock_client)
+    client = AsyncClient(api_key="test-key", insights_client=mock_client)
 
     response = await client.models.generate_content(
         model="gemini-2.5-flash",
         contents=["What's the weather in San Francisco?"],
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
     )
 
     assert response == mock_gemini_response_with_function_calls
@@ -637,12 +637,12 @@ async def test_async_cache_and_reasoning_tokens(mock_client, mock_google_genai_c
         return_value=mock_response
     )
 
-    client = AsyncClient(api_key="test-key", posthog_client=mock_client)
+    client = AsyncClient(api_key="test-key", insights_client=mock_client)
 
     response = await client.models.generate_content(
         model="gemini-2.5-pro",
         contents="Test with cache",
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
     )
 
     assert response == mock_response
@@ -689,12 +689,12 @@ async def test_async_streaming_cache_and_reasoning_tokens(
         return_value=mock_streaming_response()
     )
 
-    client = AsyncClient(api_key="test-key", posthog_client=mock_client)
+    client = AsyncClient(api_key="test-key", insights_client=mock_client)
 
     response = await client.models.generate_content_stream(
         model="gemini-2.5-pro",
         contents="Test streaming with cache",
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
     )
 
     # Consume the stream
@@ -704,7 +704,7 @@ async def test_async_streaming_cache_and_reasoning_tokens(
 
     assert len(result) == 2
 
-    # Check PostHog capture was called
+    # Check Insights capture was called
     assert mock_client.capture.call_count == 1
 
     call_args = mock_client.capture.call_args[1]
@@ -761,11 +761,11 @@ async def test_async_web_search_grounding(mock_client, mock_google_genai_client)
         return_value=mock_response
     )
 
-    client = AsyncClient(api_key="test-key", posthog_client=mock_client)
+    client = AsyncClient(api_key="test-key", insights_client=mock_client)
     response = await client.models.generate_content(
         model="gemini-2.5-flash",
         contents="What's the latest news?",
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
     )
 
     assert response == mock_response
@@ -829,12 +829,12 @@ async def test_async_streaming_with_web_search(mock_client, mock_google_genai_cl
         return_value=mock_streaming_response()
     )
 
-    client = AsyncClient(api_key="test-key", posthog_client=mock_client)
+    client = AsyncClient(api_key="test-key", insights_client=mock_client)
 
     response = await client.models.generate_content_stream(
         model="gemini-2.5-flash",
         contents="What's the latest news?",
-        posthog_distinct_id="test-id",
+        insights_distinct_id="test-id",
     )
 
     chunks = []

hanzo_insights/test/ai/langchain/test_callbacks.py

@@ -21,8 +21,8 @@ try:
     from langgraph.graph.state import END, START, StateGraph
     from langgraph.prebuilt import create_react_agent
 
-    from posthog.ai.langchain import CallbackHandler
-    from posthog.ai.langchain.callbacks import GenerationMetadata, SpanMetadata
+    from hanzo_insights.ai.langchain import CallbackHandler
+    from hanzo_insights.ai.langchain.callbacks import GenerationMetadata, SpanMetadata
 
     LANGCHAIN_AVAILABLE = True
 except ImportError:
@@ -53,9 +53,9 @@ ANTHROPIC_API_KEY = os.getenv("ANTHROPIC_API_KEY")
 
 @pytest.fixture(scope="function")
 def mock_client():
-    with patch("posthog.client.Client") as mock_client:
+    with patch("hanzo_insights.client.Client") as mock_client:
         mock_client.privacy_mode = False
-        logging.getLogger("posthog").setLevel(logging.DEBUG)
+        logging.getLogger("hanzo_insights").setLevel(logging.DEBUG)
         yield mock_client
 
 
@@ -101,7 +101,7 @@ def test_metadata_capture(mock_client):
             run_id,
             messages=[{"role": "user", "content": "Who won the world series in 2020?"}],
             invocation_params={"temperature": 0.5},
-            metadata={"ls_model_name": "hog-mini", "ls_provider": "posthog"},
+            metadata={"ls_model_name": "hog-mini", "ls_provider": "hanzo_insights"},
             name="test",
         )
     expected = GenerationMetadata(
@@ -109,11 +109,11 @@ def test_metadata_capture(mock_client):
         input=[{"role": "user", "content": "Who won the world series in 2020?"}],
         start_time=1234567890,
         model_params={"temperature": 0.5},
-        provider="posthog",
+        provider="hanzo_insights",
         base_url="https://us.posthog.com",
         name="test",
         end_time=None,
-        posthog_properties=None,
+        insights_properties=None,
     )
     assert callbacks._runs[run_id] == expected
     with patch("time.time", return_value=1234567891):
@@ -1049,7 +1049,7 @@ def test_base_url_retrieval(mock_client):
     prompt = ChatPromptTemplate.from_messages([("user", "Foo")])
     chain = prompt | ChatOpenAI(
         api_key="test",
-        model="posthog-mini",
+        model="insights-mini",
         base_url="https://test.posthog.com",
     )
     callbacks = CallbackHandler(mock_client)
@@ -1257,7 +1257,7 @@ def test_metadata_tools(mock_client):
             run_id,
             messages=[{"role": "user", "content": "What's the weather like in SF?"}],
             invocation_params={"temperature": 0.5, "tools": tools},
-            metadata={"ls_model_name": "hog-mini", "ls_provider": "posthog"},
+            metadata={"ls_model_name": "hog-mini", "ls_provider": "hanzo_insights"},
             name="test",
         )
     expected = GenerationMetadata(
@@ -1265,12 +1265,12 @@ def test_metadata_tools(mock_client):
         input=[{"role": "user", "content": "What's the weather like in SF?"}],
         start_time=1234567890,
         model_params={"temperature": 0.5},
-        provider="posthog",
+        provider="hanzo_insights",
         base_url="https://us.posthog.com",
         name="test",
         tools=tools,
         end_time=None,
-        posthog_properties=None,
+        insights_properties=None,
     )
     assert callbacks._runs[run_id] == expected
     with patch("time.time", return_value=1234567891):
@@ -1867,8 +1867,8 @@ def test_openai_reasoning_tokens_o4_mini(mock_client):
 
 
 def test_callback_handler_without_client():
-    """Test that CallbackHandler works properly when no PostHog client is passed."""
-    with patch("posthog.ai.langchain.callbacks.setup") as mock_setup:
+    """Test that CallbackHandler works properly when no Insights client is passed."""
+    with patch("hanzo_insights.ai.langchain.callbacks.setup") as mock_setup:
         mock_client = mock_setup.return_value
 
         callbacks = CallbackHandler()
@@ -1894,7 +1894,7 @@ def test_callback_handler_without_client():
 
 def test_convert_message_to_dict_tool_calls():
     """Test that _convert_message_to_dict properly converts tool calls in AIMessage."""
-    from posthog.ai.langchain.callbacks import _convert_message_to_dict
+    from hanzo_insights.ai.langchain.callbacks import _convert_message_to_dict
     from langchain_core.messages import AIMessage
     from langchain_core.messages.tool import ToolCall
 
@@ -1984,7 +1984,7 @@ def test_tool_definition(mock_client):
     assert run == expected
     assert callbacks._runs == {}
 
-    # Now test that the tools are properly captured in the PostHog event
+    # Now test that the tools are properly captured in the Insights event
     mock_response = MagicMock()
     mock_response.generations = [[MagicMock()]]
 
@@ -2236,7 +2236,7 @@ def test_agent_action_and_finish_imports():
         from langchain.schema.agent import AgentAction, AgentFinish  # type: ignore
 
     # Verify they're available in the callbacks module
-    from posthog.ai.langchain.callbacks import CallbackHandler
+    from hanzo_insights.ai.langchain.callbacks import CallbackHandler
 
     # Test on_agent_action with mock data
     mock_client = MagicMock()
@@ -2266,8 +2266,8 @@ def test_agent_action_and_finish_imports():
     assert call_args["event"] == "$ai_span"
 
 
-def test_posthog_properties_field_in_generation_metadata(mock_client):
-    """Test that posthog_properties is properly stored in GenerationMetadata."""
+def test_insights_properties_field_in_generation_metadata(mock_client):
+    """Test that insights_properties is properly stored in GenerationMetadata."""
     callbacks = CallbackHandler(mock_client)
     run_id = uuid.uuid4()
 
@@ -2281,7 +2281,7 @@ def test_posthog_properties_field_in_generation_metadata(mock_client):
             metadata={
                 "ls_model_name": "gpt-4o",
                 "ls_provider": "openai",
-                "posthog_properties": {"$ai_billable": True},
+                "insights_properties": {"$ai_billable": True},
             },
             name="test",
         )
@@ -2294,11 +2294,11 @@ def test_posthog_properties_field_in_generation_metadata(mock_client):
         provider="openai",
         base_url="https://api.openai.com",
         name="test",
-        posthog_properties={"$ai_billable": True},
+        insights_properties={"$ai_billable": True},
         end_time=None,
     )
     assert callbacks._runs[run_id] == expected
-    assert callbacks._runs[run_id].posthog_properties == {"$ai_billable": True}
+    assert callbacks._runs[run_id].insights_properties == {"$ai_billable": True}
 
     callbacks._pop_run_metadata(run_id)
 
@@ -2313,15 +2313,15 @@ def test_posthog_properties_field_in_generation_metadata(mock_client):
             metadata={
                 "ls_model_name": "gpt-4o",
                 "ls_provider": "openai",
-                "posthog_properties": {"$ai_billable": False},
+                "insights_properties": {"$ai_billable": False},
             },
             name="test",
         )
 
-    assert callbacks._runs[run_id2].posthog_properties == {"$ai_billable": False}
+    assert callbacks._runs[run_id2].insights_properties == {"$ai_billable": False}
     callbacks._pop_run_metadata(run_id2)
 
-    # Test when posthog_properties not provided
+    # Test when insights_properties not provided
     run_id3 = uuid.uuid4()
     with patch("time.time", return_value=1234567890):
         callbacks._set_llm_metadata(
@@ -2333,7 +2333,7 @@ def test_posthog_properties_field_in_generation_metadata(mock_client):
             name="test",
         )
 
-    assert callbacks._runs[run_id3].posthog_properties is None
+    assert callbacks._runs[run_id3].insights_properties is None
 
 
 def test_billable_property_in_generation_event(mock_client):
@@ -2349,7 +2349,7 @@ def test_billable_property_in_generation_event(mock_client):
             run_id,
             messages=[{"role": "user", "content": "Test"}],
             metadata={
-                "posthog_properties": {"$ai_billable": True},
+                "insights_properties": {"$ai_billable": True},
                 "ls_model_name": "test-model",
             },
             invocation_params={},
@@ -2412,12 +2412,12 @@ def test_billable_with_real_chain(mock_client):
             metadata={
                 "ls_model_name": "fake-model",
                 "ls_provider": "fake",
-                "posthog_properties": {"$ai_billable": True},
+                "insights_properties": {"$ai_billable": True},
             },
             invocation_params={"temperature": 0.7},
         )
 
-    assert callbacks._runs[run_id].posthog_properties == {"$ai_billable": True}
+    assert callbacks._runs[run_id].insights_properties == {"$ai_billable": True}
 
     mock_response = MagicMock()
     mock_response.generations = [[MagicMock()]]
@@ -2441,3 +2441,206 @@ def test_billable_with_real_chain(mock_client):
     assert props["$ai_billable"] is True
     assert props["$ai_model"] == "fake-model"
     assert props["$ai_provider"] == "fake"
+
+
+# Exception Capture Integration Tests
+
+
+def test_exception_autocapture_on_span_error():
+    """Test that capture_exception is called when a span errors and autocapture is enabled."""
+    mock_client = MagicMock()
+    mock_client.privacy_mode = False
+    mock_client.enable_exception_autocapture = True
+    mock_client.capture_exception.return_value = "exception-uuid-123"
+
+    def failing_span(_):
+        raise ValueError("test error")
+
+    callbacks = [CallbackHandler(mock_client)]
+    chain = RunnableLambda(failing_span)
+
+    try:
+        chain.invoke({}, config={"callbacks": callbacks})
+    except ValueError:
+        pass
+
+    # Verify capture_exception was called
+    assert mock_client.capture_exception.call_count == 1
+    exception_call = mock_client.capture_exception.call_args
+    assert isinstance(exception_call[0][0], ValueError)
+    assert str(exception_call[0][0]) == "test error"
+
+
+def test_exception_autocapture_adds_exception_id_to_span_event():
+    """Test that $exception_event_id is added to the span event properties."""
+    mock_client = MagicMock()
+    mock_client.privacy_mode = False
+    mock_client.enable_exception_autocapture = True
+    mock_client.capture_exception.return_value = "exception-uuid-456"
+
+    def failing_span(_):
+        raise ValueError("test error")
+
+    callbacks = [CallbackHandler(mock_client)]
+    chain = RunnableLambda(failing_span)
+
+    try:
+        chain.invoke({}, config={"callbacks": callbacks})
+    except ValueError:
+        pass
+
+    # Find the span event (should have $ai_is_error=True)
+    span_calls = [
+        call
+        for call in mock_client.capture.call_args_list
+        if call[1].get("properties", {}).get("$ai_is_error") is True
+    ]
+    assert len(span_calls) >= 1
+
+    span_props = span_calls[0][1]["properties"]
+    assert span_props["$exception_event_id"] == "exception-uuid-456"
+    assert span_props["$ai_error"] == "ValueError: test error"
+
+
+def test_exception_autocapture_disabled_does_not_capture():
+    """Test that capture_exception is NOT called when autocapture is disabled."""
+    mock_client = MagicMock()
+    mock_client.privacy_mode = False
+    mock_client.enable_exception_autocapture = False
+
+    def failing_span(_):
+        raise ValueError("test error")
+
+    callbacks = [CallbackHandler(mock_client)]
+    chain = RunnableLambda(failing_span)
+
+    try:
+        chain.invoke({}, config={"callbacks": callbacks})
+    except ValueError:
+        pass
+
+    # Verify capture_exception was NOT called
+    assert mock_client.capture_exception.call_count == 0
+
+    # But the span event should still have error info
+    span_calls = [
+        call
+        for call in mock_client.capture.call_args_list
+        if call[1].get("properties", {}).get("$ai_is_error") is True
+    ]
+    assert len(span_calls) >= 1
+
+    span_props = span_calls[0][1]["properties"]
+    assert "$exception_event_id" not in span_props
+    assert span_props["$ai_error"] == "ValueError: test error"
+
+
+def test_exception_autocapture_on_llm_generation_error(mock_client):
+    """Test that capture_exception is called when an LLM generation fails."""
+    mock_client.privacy_mode = False
+    mock_client.enable_exception_autocapture = True
+    mock_client.capture_exception.return_value = "exception-uuid-789"
+
+    callbacks = CallbackHandler(mock_client)
+    run_id = uuid.uuid4()
+
+    # Simulate LLM start
+    callbacks.on_llm_start(
+        serialized={"kwargs": {"openai_api_base": "https://api.openai.com"}},
+        prompts=["Hello"],
+        run_id=run_id,
+    )
+
+    # Simulate LLM error
+    error = Exception("API rate limit exceeded")
+    callbacks.on_llm_error(error, run_id=run_id)
+
+    # Verify capture_exception was called
+    assert mock_client.capture_exception.call_count == 1
+    exception_call = mock_client.capture_exception.call_args
+    assert exception_call[0][0] is error
+
+    # Verify the generation event has $exception_event_id
+    generation_calls = [
+        call
+        for call in mock_client.capture.call_args_list
+        if call[1].get("event") == "$ai_generation"
+    ]
+    assert len(generation_calls) == 1
+
+    gen_props = generation_calls[0][1]["properties"]
+    assert gen_props["$exception_event_id"] == "exception-uuid-789"
+    assert gen_props["$ai_is_error"] is True
+
+
+def test_exception_autocapture_passes_ai_properties_to_exception():
+    """Test that AI properties are passed to the exception event."""
+    mock_client = MagicMock()
+    mock_client.privacy_mode = False
+    mock_client.enable_exception_autocapture = True
+    mock_client.capture_exception.return_value = "exception-uuid-abc"
+
+    callbacks = CallbackHandler(
+        mock_client,
+        distinct_id="user-123",
+        properties={"custom_prop": "custom_value"},
+    )
+    run_id = uuid.uuid4()
+
+    # Simulate LLM start
+    callbacks.on_llm_start(
+        serialized={"kwargs": {"openai_api_base": "https://api.openai.com"}},
+        prompts=["Hello"],
+        run_id=run_id,
+    )
+
+    # Simulate LLM error
+    error = Exception("API error")
+    callbacks.on_llm_error(error, run_id=run_id)
+
+    # Verify capture_exception received the properties
+    exception_call = mock_client.capture_exception.call_args
+    props = exception_call[1]["properties"]
+
+    # Should have AI-related properties
+    assert "$ai_trace_id" in props
+    assert "$ai_is_error" in props
+    assert props["$ai_is_error"] is True
+
+    # Should have distinct_id passed through
+    assert exception_call[1]["distinct_id"] == "user-123"
+
+
+def test_exception_autocapture_none_return_no_exception_id():
+    """Test that when capture_exception returns None, no $exception_event_id is added."""
+    mock_client = MagicMock()
+    mock_client.privacy_mode = False
+    mock_client.enable_exception_autocapture = True
+    mock_client.capture_exception.return_value = (
+        None  # e.g., exception already captured
+    )
+
+    def failing_span(_):
+        raise ValueError("test error")
+
+    callbacks = [CallbackHandler(mock_client)]
+    chain = RunnableLambda(failing_span)
+
+    try:
+        chain.invoke({}, config={"callbacks": callbacks})
+    except ValueError:
+        pass
+
+    # capture_exception was called but returned None
+    assert mock_client.capture_exception.call_count == 1
+
+    # Span event should NOT have $exception_event_id
+    span_calls = [
+        call
+        for call in mock_client.capture.call_args_list
+        if call[1].get("properties", {}).get("$ai_is_error") is True
+    ]
+    assert len(span_calls) >= 1
+
+    span_props = span_calls[0][1]["properties"]
+    assert "$exception_event_id" not in span_props

hanzo_insights/test/ai/openai/test_openai.py

@@ -1,3 +1,4 @@
+import json
 import time
 from unittest.mock import AsyncMock, patch
 
@@ -33,8 +34,8 @@ try:
         ParsedResponseOutputText,
     )
 
-    from posthog.ai.openai import OpenAI
-    from posthog.ai.openai.openai_async import AsyncOpenAI
+    from hanzo_insights.ai.openai import OpenAI
+    from hanzo_insights.ai.openai.openai_async import AsyncOpenAI
 
     OPENAI_AVAILABLE = True
 except ImportError:
@@ -48,7 +49,7 @@ pytestmark = pytest.mark.skipif(
 
 @pytest.fixture
 def mock_client():
-    with patch("posthog.client.Client") as mock_client:
+    with patch("hanzo_insights.client.Client") as mock_client:
         mock_client.privacy_mode = False
         yield mock_client
 
@@ -466,12 +467,12 @@ def test_basic_completion(mock_client, mock_openai_response):
         "openai.resources.chat.completions.Completions.create",
         return_value=mock_openai_response,
     ):
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
         response = client.chat.completions.create(
             model="gpt-4",
             messages=[{"role": "user", "content": "Hello"}],
-            posthog_distinct_id="test-id",
-            posthog_properties={"foo": "bar"},
+            insights_distinct_id="test-id",
+            insights_properties={"foo": "bar"},
         )
 
         assert response == mock_openai_response
@@ -496,6 +497,15 @@ def test_basic_completion(mock_client, mock_openai_response):
         assert props["$ai_http_status"] == 200
         assert props["foo"] == "bar"
         assert isinstance(props["$ai_latency"], float)
+        # Verify raw usage metadata is passed for backend processing
+        assert "$ai_usage" in props
+        assert props["$ai_usage"] is not None
+        # Verify it's JSON-serializable
+        json.dumps(props["$ai_usage"])
+        # Verify it has expected structure
+        assert isinstance(props["$ai_usage"], dict)
+        assert "prompt_tokens" in props["$ai_usage"]
+        assert "completion_tokens" in props["$ai_usage"]
 
 
 def test_embeddings(mock_client, mock_embedding_response):
@@ -503,12 +513,12 @@ def test_embeddings(mock_client, mock_embedding_response):
         "openai.resources.embeddings.Embeddings.create",
         return_value=mock_embedding_response,
     ):
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
         response = client.embeddings.create(
             model="text-embedding-3-small",
             input="Hello world",
-            posthog_distinct_id="test-id",
-            posthog_properties={"foo": "bar"},
+            insights_distinct_id="test-id",
+            insights_properties={"foo": "bar"},
         )
 
         assert response == mock_embedding_response
@@ -533,12 +543,12 @@ def test_groups(mock_client, mock_openai_response):
         "openai.resources.chat.completions.Completions.create",
         return_value=mock_openai_response,
     ):
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
         response = client.chat.completions.create(
             model="gpt-4",
             messages=[{"role": "user", "content": "Hello"}],
-            posthog_distinct_id="test-id",
-            posthog_groups={"company": "test_company"},
+            insights_distinct_id="test-id",
+            insights_groups={"company": "test_company"},
         )
 
         assert response == mock_openai_response
@@ -554,12 +564,12 @@ def test_privacy_mode_local(mock_client, mock_openai_response):
         "openai.resources.chat.completions.Completions.create",
         return_value=mock_openai_response,
     ):
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
         response = client.chat.completions.create(
             model="gpt-4",
             messages=[{"role": "user", "content": "Hello"}],
-            posthog_distinct_id="test-id",
-            posthog_privacy_mode=True,
+            insights_distinct_id="test-id",
+            insights_privacy_mode=True,
         )
 
         assert response == mock_openai_response
@@ -577,12 +587,12 @@ def test_privacy_mode_global(mock_client, mock_openai_response):
         return_value=mock_openai_response,
     ):
         mock_client.privacy_mode = True
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
         response = client.chat.completions.create(
             model="gpt-4",
             messages=[{"role": "user", "content": "Hello"}],
-            posthog_distinct_id="test-id",
-            posthog_privacy_mode=False,
+            insights_distinct_id="test-id",
+            insights_privacy_mode=False,
         )
 
         assert response == mock_openai_response
@@ -599,7 +609,7 @@ def test_error(mock_client, mock_openai_response):
         "openai.resources.chat.completions.Completions.create",
         side_effect=Exception("Test error"),
     ):
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
         with pytest.raises(Exception):
             client.chat.completions.create(
                 model="gpt-4", messages=[{"role": "user", "content": "Hello"}]
@@ -618,12 +628,12 @@ def test_cached_tokens(mock_client, mock_openai_response_with_cached_tokens):
         "openai.resources.chat.completions.Completions.create",
         return_value=mock_openai_response_with_cached_tokens,
     ):
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
         response = client.chat.completions.create(
             model="gpt-4",
             messages=[{"role": "user", "content": "Hello"}],
-            posthog_distinct_id="test-id",
-            posthog_properties={"foo": "bar"},
+            insights_distinct_id="test-id",
+            insights_properties={"foo": "bar"},
         )
 
         assert response == mock_openai_response_with_cached_tokens
@@ -656,7 +666,7 @@ def test_tool_calls(mock_client, mock_openai_response_with_tool_calls):
         "openai.resources.chat.completions.Completions.create",
         return_value=mock_openai_response_with_tool_calls,
     ):
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
         response = client.chat.completions.create(
             model="gpt-4",
             messages=[
@@ -672,7 +682,7 @@ def test_tool_calls(mock_client, mock_openai_response_with_tool_calls):
                     },
                 }
             ],
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         assert response == mock_openai_response_with_tool_calls
@@ -729,7 +739,7 @@ def test_tool_calls_only_no_content(mock_client, mock_openai_response_tool_calls
         "openai.resources.chat.completions.Completions.create",
         return_value=mock_openai_response_tool_calls_only,
     ):
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
         response = client.chat.completions.create(
             model="gpt-4",
             messages=[{"role": "user", "content": "Get weather for New York"}],
@@ -743,7 +753,7 @@ def test_tool_calls_only_no_content(mock_client, mock_openai_response_tool_calls
                     },
                 }
             ],
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         assert response == mock_openai_response_tool_calls_only
@@ -783,7 +793,7 @@ def test_responses_api_tool_calls(mock_client, mock_responses_api_with_tool_call
         "openai.resources.responses.Responses.create",
         return_value=mock_responses_api_with_tool_calls,
     ):
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
         response = client.responses.create(
             model="gpt-4o-mini",
             input=[{"role": "user", "content": "What's the weather in Chicago?"}],
@@ -798,7 +808,7 @@ def test_responses_api_tool_calls(mock_client, mock_responses_api_with_tool_call
                     },
                 }
             ],
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         assert response == mock_responses_api_with_tool_calls
@@ -841,7 +851,7 @@ def test_streaming_with_tool_calls(mock_client, streaming_tool_call_chunks):
         # Set up the mock to return our chunks when iterated
         mock_create.return_value = streaming_tool_call_chunks
 
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
 
         # Call the streaming method
         response_generator = client.chat.completions.create(
@@ -860,7 +870,7 @@ def test_streaming_with_tool_calls(mock_client, streaming_tool_call_chunks):
                 }
             ],
             stream=True,
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         # Consume the generator to trigger the event capture
@@ -922,6 +932,16 @@ def test_streaming_with_tool_calls(mock_client, streaming_tool_call_chunks):
         assert props["$ai_input_tokens"] == 20
         assert props["$ai_output_tokens"] == 15
 
+        # Verify raw usage is captured in streaming mode
+        assert "$ai_usage" in props
+        assert props["$ai_usage"] is not None
+        # Verify it's JSON-serializable
+        json.dumps(props["$ai_usage"])
+        # Verify it has expected structure (merged from chunks)
+        assert isinstance(props["$ai_usage"], dict)
+        assert "prompt_tokens" in props["$ai_usage"]
+        assert "completion_tokens" in props["$ai_usage"]
+
 
 # test responses api
 def test_responses_api(mock_client, mock_openai_response_with_responses_api):
@@ -929,12 +949,12 @@ def test_responses_api(mock_client, mock_openai_response_with_responses_api):
         "openai.resources.responses.Responses.create",
         return_value=mock_openai_response_with_responses_api,
     ):
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
         response = client.responses.create(
             model="gpt-4o-mini",
             input="Hello",
-            posthog_distinct_id="test-id",
-            posthog_properties={"foo": "bar"},
+            insights_distinct_id="test-id",
+            insights_properties={"foo": "bar"},
         )
         assert response == mock_openai_response_with_responses_api
         assert mock_client.capture.call_count == 1
@@ -966,7 +986,7 @@ def test_responses_parse(mock_client, mock_parsed_response):
         "openai.resources.responses.Responses.parse",
         return_value=mock_parsed_response,
     ):
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
         response = client.responses.parse(
             model="gpt-4o-2024-08-06",
             input=[
@@ -996,8 +1016,8 @@ def test_responses_parse(mock_client, mock_parsed_response):
                     },
                 }
             },
-            posthog_distinct_id="test-id",
-            posthog_properties={"foo": "bar"},
+            insights_distinct_id="test-id",
+            insights_properties={"foo": "bar"},
         )
 
         assert response == mock_parsed_response
@@ -1081,15 +1101,15 @@ def test_responses_api_streaming_with_tokens(mock_client):
         "openai.resources.responses.Responses.create",
         side_effect=mock_streaming_response,
     ):
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
 
         # Consume the streaming response
         response = client.responses.create(
             model="gpt-4o-mini",
             input=[{"role": "user", "content": "Test message"}],
             stream=True,
-            posthog_distinct_id="test-id",
-            posthog_properties={"test": "streaming"},
+            insights_distinct_id="test-id",
+            insights_properties={"test": "streaming"},
         )
 
         # Consume all chunks
@@ -1133,7 +1153,7 @@ async def test_async_chat_streaming_with_tool_calls(
     with patch(
         "openai.resources.chat.completions.AsyncCompletions.create", new=mock_create
     ):
-        client = AsyncOpenAI(api_key="test-key", posthog_client=mock_client)
+        client = AsyncOpenAI(api_key="test-key", insights_client=mock_client)
 
         response_stream = await client.chat.completions.create(
             model="gpt-4",
@@ -1151,7 +1171,7 @@ async def test_async_chat_streaming_with_tool_calls(
                 }
             ],
             stream=True,
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         chunks = []
@@ -1219,14 +1239,14 @@ async def test_async_responses_streaming_with_tokens(mock_client):
         return chunk_iterable()
 
     with patch("openai.resources.responses.AsyncResponses.create", new=mock_create):
-        client = AsyncOpenAI(api_key="test-key", posthog_client=mock_client)
+        client = AsyncOpenAI(api_key="test-key", insights_client=mock_client)
 
         response_stream = await client.responses.create(
             model="gpt-4o-mini",
             input=[{"role": "user", "content": "Test message"}],
             stream=True,
-            posthog_distinct_id="test-id",
-            posthog_properties={"test": "streaming"},
+            insights_distinct_id="test-id",
+            insights_properties={"test": "streaming"},
         )
 
         async for _ in response_stream:
@@ -1254,13 +1274,13 @@ async def test_async_embeddings_create(mock_client, mock_embedding_response):
     mock_create = AsyncMock(return_value=mock_embedding_response)
 
     with patch("openai.resources.embeddings.AsyncEmbeddings.create", new=mock_create):
-        client = AsyncOpenAI(api_key="test-key", posthog_client=mock_client)
+        client = AsyncOpenAI(api_key="test-key", insights_client=mock_client)
 
         response = await client.embeddings.create(
             model="text-embedding-3-small",
             input="Hello world",
-            posthog_distinct_id="test-id",
-            posthog_properties={"foo": "bar"},
+            insights_distinct_id="test-id",
+            insights_properties={"foo": "bar"},
         )
 
     assert response == mock_embedding_response
@@ -1284,7 +1304,7 @@ def test_tool_definition(mock_client, mock_openai_response):
         "openai.resources.chat.completions.Completions.create",
         return_value=mock_openai_response,
     ):
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
 
         # Define tools to be passed to the create function
         tools = [
@@ -1311,8 +1331,8 @@ def test_tool_definition(mock_client, mock_openai_response):
             model="gpt-4o-mini",
             messages=[{"role": "user", "content": "hey"}],
             tools=tools,
-            posthog_distinct_id="test-id",
-            posthog_properties={"foo": "bar"},
+            insights_distinct_id="test-id",
+            insights_properties={"foo": "bar"},
         )
 
         assert response == mock_openai_response
@@ -1372,11 +1392,11 @@ def test_web_search_perplexity_style(mock_client):
     mock_response = MockResponseWithAnnotations()
 
     with patch("openai.resources.chat.Completions.create", return_value=mock_response):
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
         response = client.chat.completions.create(
             model="gpt-4-turbo",
             messages=[{"role": "user", "content": "What's happening in tech?"}],
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         assert response == mock_response
@@ -1419,19 +1439,19 @@ def test_web_search_responses_api(mock_client):
         "openai.resources.responses.Responses.create", return_value=mock_response
     ):
         # Manually call the tracking since we're testing the converter logic
-        from posthog.ai.utils import call_llm_and_track_usage
+        from hanzo_insights.ai.utils import call_llm_and_track_usage
 
         def mock_create_call(**kwargs):
             return mock_response
 
         result = call_llm_and_track_usage(
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
             ph_client=mock_client,
             provider="openai",
-            posthog_trace_id=None,
-            posthog_properties=None,
-            posthog_privacy_mode=False,
-            posthog_groups=None,
+            insights_trace_id=None,
+            insights_properties=None,
+            insights_privacy_mode=False,
+            insights_groups=None,
             base_url="https://api.openai.com/v1",
             call_method=mock_create_call,
             model="gpt-4o",
@@ -1513,12 +1533,12 @@ def test_streaming_with_web_search(mock_client, streaming_web_search_chunks):
     with patch("openai.resources.chat.completions.Completions.create") as mock_create:
         mock_create.return_value = streaming_web_search_chunks
 
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
         response_generator = client.chat.completions.create(
             model="gpt-4",
             messages=[{"role": "user", "content": "Search for recent news"}],
             stream=True,
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         # Consume the generator to trigger the event capture
@@ -1549,12 +1569,12 @@ def test_streaming_with_web_search_on_non_usage_chunk(
     with patch("openai.resources.chat.completions.Completions.create") as mock_create:
         mock_create.return_value = streaming_web_search_chunks
 
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
         response_generator = client.chat.completions.create(
             model="gpt-4",
             messages=[{"role": "user", "content": "Search for recent news"}],
             stream=True,
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         # Consume the generator to trigger the event capture
@@ -1609,12 +1629,12 @@ async def test_async_chat_with_web_search(mock_client):
     with patch(
         "openai.resources.chat.completions.AsyncCompletions.create", new=mock_create
     ):
-        client = AsyncOpenAI(api_key="test-key", posthog_client=mock_client)
+        client = AsyncOpenAI(api_key="test-key", insights_client=mock_client)
 
         response = await client.chat.completions.create(
             model="gpt-4",
             messages=[{"role": "user", "content": "Search for recent news"}],
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
     assert response == mock_response
@@ -1652,13 +1672,13 @@ async def test_async_chat_streaming_with_web_search(
     with patch(
         "openai.resources.chat.completions.AsyncCompletions.create", new=mock_create
     ):
-        client = AsyncOpenAI(api_key="test-key", posthog_client=mock_client)
+        client = AsyncOpenAI(api_key="test-key", insights_client=mock_client)
 
         response_stream = await client.chat.completions.create(
             model="gpt-4",
             messages=[{"role": "user", "content": "Search for recent news"}],
             stream=True,
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         chunks = []
@@ -1722,13 +1742,13 @@ def test_streaming_chat_extracts_model_from_chunk_when_not_in_kwargs(mock_client
     with patch("openai.resources.chat.completions.Completions.create") as mock_create:
         mock_create.return_value = chunks
 
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
 
         # Note: NOT passing model in kwargs - simulates stored prompt usage
         response_generator = client.chat.completions.create(
             messages=[{"role": "user", "content": "Hello"}],
             stream=True,
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         # Consume the generator
@@ -1768,13 +1788,13 @@ def test_streaming_chat_prefers_kwargs_model_over_chunk_model(mock_client):
     with patch("openai.resources.chat.completions.Completions.create") as mock_create:
         mock_create.return_value = chunks
 
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
 
         response_generator = client.chat.completions.create(
             model="gpt-4o-from-kwargs",  # Explicitly passed model
             messages=[{"role": "user", "content": "Hello"}],
             stream=True,
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         list(response_generator)
@@ -1819,13 +1839,13 @@ def test_streaming_responses_api_extracts_model_from_response_object(mock_client
     with patch("openai.resources.responses.Responses.create") as mock_create:
         mock_create.return_value = iter(chunks)
 
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
 
         # Note: NOT passing model - simulates stored prompt
         response_generator = client.responses.create(
             input=[{"role": "user", "content": "Hello"}],
             stream=True,
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         list(response_generator)
@@ -1866,12 +1886,12 @@ def test_non_streaming_extracts_model_from_response(mock_client):
         "openai.resources.chat.completions.Completions.create",
         return_value=mock_response,
     ):
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
 
         # Note: NOT passing model in kwargs
         response = client.chat.completions.create(
             messages=[{"role": "user", "content": "Hello"}],
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         assert response == mock_response
@@ -1928,12 +1948,12 @@ def test_non_streaming_responses_api_extracts_model_from_response(mock_client):
         "openai.resources.responses.Responses.create",
         return_value=mock_response,
     ):
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
 
         # Note: NOT passing model in kwargs
         response = client.responses.create(
             input="Hello",
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         assert response == mock_response
@@ -1975,12 +1995,12 @@ def test_non_streaming_returns_none_when_no_model(mock_client):
         "openai.resources.chat.completions.Completions.create",
         return_value=mock_response,
     ):
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
 
         # Note: NOT passing model in kwargs and response has no model
         client.chat.completions.create(
             messages=[{"role": "user", "content": "Hello"}],
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         call_args = mock_client.capture.call_args[1]
@@ -2012,12 +2032,12 @@ def test_streaming_falls_back_to_unknown_when_no_model(mock_client):
     with patch("openai.resources.chat.completions.Completions.create") as mock_create:
         mock_create.return_value = [chunk]
 
-        client = OpenAI(api_key="test-key", posthog_client=mock_client)
+        client = OpenAI(api_key="test-key", insights_client=mock_client)
 
         response_generator = client.chat.completions.create(
             messages=[{"role": "user", "content": "Hello"}],
             stream=True,
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         list(response_generator)
@@ -2063,13 +2083,13 @@ async def test_async_streaming_chat_extracts_model_from_chunk(mock_client):
     with patch(
         "openai.resources.chat.completions.AsyncCompletions.create", new=mock_create
     ):
-        client = AsyncOpenAI(api_key="test-key", posthog_client=mock_client)
+        client = AsyncOpenAI(api_key="test-key", insights_client=mock_client)
 
         # Note: NOT passing model
         response_stream = await client.chat.completions.create(
             messages=[{"role": "user", "content": "Hello"}],
             stream=True,
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         async for _ in response_stream:
@@ -2117,12 +2137,12 @@ async def test_async_streaming_responses_extracts_model_from_response(mock_clien
         return chunk_iterable()
 
     with patch("openai.resources.responses.AsyncResponses.create", new=mock_create):
-        client = AsyncOpenAI(api_key="test-key", posthog_client=mock_client)
+        client = AsyncOpenAI(api_key="test-key", insights_client=mock_client)
 
         response_stream = await client.responses.create(
             input=[{"role": "user", "content": "Hello"}],
             stream=True,
-            posthog_distinct_id="test-id",
+            insights_distinct_id="test-id",
         )
 
         async for _ in response_stream:

hanzo_insights/test/ai/openai_agents/__init__.py

@@ -0,0 +1 @@
+# Tests for OpenAI Agents SDK integration

hanzo_insights/test/ai/openai_agents/test_processor.py

@@ -0,0 +1,810 @@
+import logging
+from unittest.mock import MagicMock, patch
+
+import pytest
+
+try:
+    from agents.tracing.span_data import (
+        AgentSpanData,
+        CustomSpanData,
+        FunctionSpanData,
+        GenerationSpanData,
+        GuardrailSpanData,
+        HandoffSpanData,
+        ResponseSpanData,
+        SpeechSpanData,
+        TranscriptionSpanData,
+    )
+
+    from hanzo_insights.ai.openai_agents import InsightsTracingProcessor, instrument
+
+    OPENAI_AGENTS_AVAILABLE = True
+except ImportError:
+    OPENAI_AGENTS_AVAILABLE = False
+
+
+# Skip all tests if OpenAI Agents SDK is not available
+pytestmark = pytest.mark.skipif(
+    not OPENAI_AGENTS_AVAILABLE, reason="OpenAI Agents SDK is not available"
+)
+
+
+@pytest.fixture(scope="function")
+def mock_client():
+    client = MagicMock()
+    client.privacy_mode = False
+    logging.getLogger("hanzo_insights").setLevel(logging.DEBUG)
+    return client
+
+
+@pytest.fixture(scope="function")
+def processor(mock_client):
+    return InsightsTracingProcessor(
+        client=mock_client,
+        distinct_id="test-user",
+        privacy_mode=False,
+    )
+
+
+@pytest.fixture
+def mock_trace():
+    trace = MagicMock()
+    trace.trace_id = "trace_123456789"
+    trace.name = "Test Workflow"
+    trace.group_id = "group_123"
+    trace.metadata = {"key": "value"}
+    return trace
+
+
+@pytest.fixture
+def mock_span():
+    span = MagicMock()
+    span.trace_id = "trace_123456789"
+    span.span_id = "span_987654321"
+    span.parent_id = None
+    span.started_at = "2024-01-01T00:00:00Z"
+    span.ended_at = "2024-01-01T00:00:01Z"
+    span.error = None
+    return span
+
+
+class TestInsightsTracingProcessor:
+    """Tests for the InsightsTracingProcessor class."""
+
+    def test_initialization(self, mock_client):
+        """Test processor initializes correctly."""
+        processor = InsightsTracingProcessor(
+            client=mock_client,
+            distinct_id="user@example.com",
+            privacy_mode=True,
+            groups={"company": "acme"},
+            properties={"env": "test"},
+        )
+
+        assert processor._client == mock_client
+        assert processor._distinct_id == "user@example.com"
+        assert processor._privacy_mode is True
+        assert processor._groups == {"company": "acme"}
+        assert processor._properties == {"env": "test"}
+
+    def test_initialization_with_callable_distinct_id(self, mock_client, mock_trace):
+        """Test processor with callable distinct_id resolver."""
+
+        def resolver(trace):
+            return trace.metadata.get("user_id", "default")
+
+        processor = InsightsTracingProcessor(
+            client=mock_client,
+            distinct_id=resolver,
+        )
+
+        mock_trace.metadata = {"user_id": "resolved-user"}
+        distinct_id = processor._get_distinct_id(mock_trace)
+        assert distinct_id == "resolved-user"
+
+    def test_on_trace_start_stores_metadata(self, processor, mock_client, mock_trace):
+        """Test that on_trace_start stores metadata but does not capture an event."""
+        processor.on_trace_start(mock_trace)
+
+        mock_client.capture.assert_not_called()
+        assert mock_trace.trace_id in processor._trace_metadata
+
+    def test_on_trace_end_captures_ai_trace(self, processor, mock_client, mock_trace):
+        """Test that on_trace_end captures $ai_trace event."""
+        processor.on_trace_start(mock_trace)
+        processor.on_trace_end(mock_trace)
+
+        mock_client.capture.assert_called_once()
+        call_kwargs = mock_client.capture.call_args[1]
+
+        assert call_kwargs["event"] == "$ai_trace"
+        assert call_kwargs["distinct_id"] == "test-user"
+        assert call_kwargs["properties"]["$ai_trace_id"] == "trace_123456789"
+        assert call_kwargs["properties"]["$ai_trace_name"] == "Test Workflow"
+        assert call_kwargs["properties"]["$ai_provider"] == "openai"
+        assert call_kwargs["properties"]["$ai_framework"] == "openai-agents"
+        assert "$ai_latency" in call_kwargs["properties"]
+
+    def test_personless_mode_when_no_distinct_id(self, mock_client, mock_trace):
+        """Test that trace events use personless mode when no distinct_id is provided."""
+        processor = InsightsTracingProcessor(
+            client=mock_client,
+        )
+
+        processor.on_trace_start(mock_trace)
+        processor.on_trace_end(mock_trace)
+
+        call_kwargs = mock_client.capture.call_args[1]
+        assert call_kwargs["properties"]["$process_person_profile"] is False
+        # Should fallback to trace_id as the distinct_id
+        assert call_kwargs["distinct_id"] == mock_trace.trace_id
+
+    def test_personless_mode_for_spans_when_no_distinct_id(
+        self, mock_client, mock_trace, mock_span
+    ):
+        """Test that span events use personless mode when no distinct_id is provided."""
+        processor = InsightsTracingProcessor(
+            client=mock_client,
+        )
+
+        processor.on_trace_start(mock_trace)
+        mock_client.capture.reset_mock()
+
+        span_data = GenerationSpanData(model="gpt-4o")
+        mock_span.span_data = span_data
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+        assert call_kwargs["properties"]["$process_person_profile"] is False
+        assert call_kwargs["distinct_id"] == mock_span.trace_id
+
+    def test_personless_mode_when_callable_returns_none(
+        self, mock_client, mock_trace, mock_span
+    ):
+        """Test personless mode when callable distinct_id returns None."""
+
+        def resolver(trace):
+            return None  # Simulate no user ID available
+
+        processor = InsightsTracingProcessor(
+            client=mock_client,
+            distinct_id=resolver,
+        )
+
+        processor.on_trace_start(mock_trace)
+        mock_client.capture.reset_mock()
+
+        span_data = GenerationSpanData(model="gpt-4o")
+        mock_span.span_data = span_data
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+        assert call_kwargs["properties"]["$process_person_profile"] is False
+        assert call_kwargs["distinct_id"] == mock_span.trace_id
+
+    def test_person_profile_when_distinct_id_provided(self, mock_client, mock_trace):
+        """Test that events create person profiles when distinct_id is provided."""
+        processor = InsightsTracingProcessor(
+            client=mock_client,
+            distinct_id="real-user",
+        )
+
+        processor.on_trace_start(mock_trace)
+        processor.on_trace_end(mock_trace)
+
+        call_kwargs = mock_client.capture.call_args[1]
+        assert "$process_person_profile" not in call_kwargs["properties"]
+
+    def test_on_trace_end_clears_metadata(self, processor, mock_client, mock_trace):
+        """Test that on_trace_end clears stored trace metadata."""
+        processor.on_trace_start(mock_trace)
+        assert mock_trace.trace_id in processor._trace_metadata
+
+        processor.on_trace_end(mock_trace)
+        assert mock_trace.trace_id not in processor._trace_metadata
+        # Also verify it captured the event
+        mock_client.capture.assert_called_once()
+
+    def test_on_span_start_tracks_time(self, processor, mock_span):
+        """Test that on_span_start records start time."""
+        processor.on_span_start(mock_span)
+        assert mock_span.span_id in processor._span_start_times
+
+    def test_generation_span_mapping(self, processor, mock_client, mock_span):
+        """Test GenerationSpanData maps to $ai_generation event."""
+        span_data = GenerationSpanData(
+            input=[{"role": "user", "content": "Hello"}],
+            output=[{"role": "assistant", "content": "Hi there!"}],
+            model="gpt-4o",
+            model_config={"temperature": 0.7, "max_tokens": 100},
+            usage={"input_tokens": 10, "output_tokens": 20},
+        )
+        mock_span.span_data = span_data
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        mock_client.capture.assert_called_once()
+        call_kwargs = mock_client.capture.call_args[1]
+
+        assert call_kwargs["event"] == "$ai_generation"
+        assert call_kwargs["properties"]["$ai_trace_id"] == "trace_123456789"
+        assert call_kwargs["properties"]["$ai_span_id"] == "span_987654321"
+        assert call_kwargs["properties"]["$ai_provider"] == "openai"
+        assert call_kwargs["properties"]["$ai_framework"] == "openai-agents"
+        assert call_kwargs["properties"]["$ai_model"] == "gpt-4o"
+        assert call_kwargs["properties"]["$ai_input_tokens"] == 10
+        assert call_kwargs["properties"]["$ai_output_tokens"] == 20
+        assert call_kwargs["properties"]["$ai_input"] == [
+            {"role": "user", "content": "Hello"}
+        ]
+        assert call_kwargs["properties"]["$ai_output_choices"] == [
+            {"role": "assistant", "content": "Hi there!"}
+        ]
+
+    def test_generation_span_with_reasoning_tokens(
+        self, processor, mock_client, mock_span
+    ):
+        """Test GenerationSpanData includes reasoning tokens when present."""
+        span_data = GenerationSpanData(
+            model="o1-preview",
+            usage={
+                "input_tokens": 100,
+                "output_tokens": 500,
+                "reasoning_tokens": 400,
+            },
+        )
+        mock_span.span_data = span_data
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+        assert call_kwargs["properties"]["$ai_reasoning_tokens"] == 400
+
+    def test_function_span_mapping(self, processor, mock_client, mock_span):
+        """Test FunctionSpanData maps to $ai_span event with type=tool."""
+        span_data = FunctionSpanData(
+            name="get_weather",
+            input='{"city": "San Francisco"}',
+            output="Sunny, 72F",
+        )
+        mock_span.span_data = span_data
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+
+        assert call_kwargs["event"] == "$ai_span"
+        assert call_kwargs["properties"]["$ai_span_name"] == "get_weather"
+        assert call_kwargs["properties"]["$ai_span_type"] == "tool"
+        assert (
+            call_kwargs["properties"]["$ai_input_state"] == '{"city": "San Francisco"}'
+        )
+        assert call_kwargs["properties"]["$ai_output_state"] == "Sunny, 72F"
+
+    def test_agent_span_mapping(self, processor, mock_client, mock_span):
+        """Test AgentSpanData maps to $ai_span event with type=agent."""
+        span_data = AgentSpanData(
+            name="CustomerServiceAgent",
+            handoffs=["TechnicalAgent", "BillingAgent"],
+            tools=["search", "get_order"],
+            output_type="str",
+        )
+        mock_span.span_data = span_data
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+
+        assert call_kwargs["event"] == "$ai_span"
+        assert call_kwargs["properties"]["$ai_span_name"] == "CustomerServiceAgent"
+        assert call_kwargs["properties"]["$ai_span_type"] == "agent"
+        assert call_kwargs["properties"]["$ai_agent_handoffs"] == [
+            "TechnicalAgent",
+            "BillingAgent",
+        ]
+        assert call_kwargs["properties"]["$ai_agent_tools"] == ["search", "get_order"]
+
+    def test_handoff_span_mapping(self, processor, mock_client, mock_span):
+        """Test HandoffSpanData maps to $ai_span event with type=handoff."""
+        span_data = HandoffSpanData(
+            from_agent="TriageAgent",
+            to_agent="TechnicalAgent",
+        )
+        mock_span.span_data = span_data
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+
+        assert call_kwargs["event"] == "$ai_span"
+        assert call_kwargs["properties"]["$ai_span_type"] == "handoff"
+        assert call_kwargs["properties"]["$ai_handoff_from_agent"] == "TriageAgent"
+        assert call_kwargs["properties"]["$ai_handoff_to_agent"] == "TechnicalAgent"
+        assert (
+            call_kwargs["properties"]["$ai_span_name"]
+            == "TriageAgent -> TechnicalAgent"
+        )
+
+    def test_guardrail_span_mapping(self, processor, mock_client, mock_span):
+        """Test GuardrailSpanData maps to $ai_span event with type=guardrail."""
+        span_data = GuardrailSpanData(
+            name="ContentFilter",
+            triggered=True,
+        )
+        mock_span.span_data = span_data
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+
+        assert call_kwargs["event"] == "$ai_span"
+        assert call_kwargs["properties"]["$ai_span_name"] == "ContentFilter"
+        assert call_kwargs["properties"]["$ai_span_type"] == "guardrail"
+        assert call_kwargs["properties"]["$ai_guardrail_triggered"] is True
+
+    def test_custom_span_mapping(self, processor, mock_client, mock_span):
+        """Test CustomSpanData maps to $ai_span event with type=custom."""
+        span_data = CustomSpanData(
+            name="database_query",
+            data={"query": "SELECT * FROM users", "rows": 100},
+        )
+        mock_span.span_data = span_data
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+
+        assert call_kwargs["event"] == "$ai_span"
+        assert call_kwargs["properties"]["$ai_span_name"] == "database_query"
+        assert call_kwargs["properties"]["$ai_span_type"] == "custom"
+        assert call_kwargs["properties"]["$ai_custom_data"] == {
+            "query": "SELECT * FROM users",
+            "rows": 100,
+        }
+
+    def test_privacy_mode_redacts_content(self, mock_client, mock_span):
+        """Test that privacy_mode redacts input/output content."""
+        processor = InsightsTracingProcessor(
+            client=mock_client,
+            distinct_id="test-user",
+            privacy_mode=True,
+        )
+
+        span_data = GenerationSpanData(
+            input=[{"role": "user", "content": "Secret message"}],
+            output=[{"role": "assistant", "content": "Secret response"}],
+            model="gpt-4o",
+            usage={"input_tokens": 10, "output_tokens": 20},
+        )
+        mock_span.span_data = span_data
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+
+        # Content should be redacted
+        assert call_kwargs["properties"]["$ai_input"] is None
+        assert call_kwargs["properties"]["$ai_output_choices"] is None
+        # Token counts should still be present
+        assert call_kwargs["properties"]["$ai_input_tokens"] == 10
+        assert call_kwargs["properties"]["$ai_output_tokens"] == 20
+
+    def test_error_handling_in_span(self, processor, mock_client, mock_span):
+        """Test that span errors are captured correctly."""
+        span_data = GenerationSpanData(model="gpt-4o")
+        mock_span.span_data = span_data
+        mock_span.error = {"message": "Rate limit exceeded", "data": {"code": 429}}
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+
+        assert call_kwargs["properties"]["$ai_is_error"] is True
+        assert call_kwargs["properties"]["$ai_error"] == "Rate limit exceeded"
+
+    def test_generation_span_includes_total_tokens(
+        self, processor, mock_client, mock_span
+    ):
+        """Test that $ai_total_tokens is calculated and included."""
+        span_data = GenerationSpanData(
+            model="gpt-4o",
+            usage={"input_tokens": 100, "output_tokens": 50},
+        )
+        mock_span.span_data = span_data
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+        assert call_kwargs["properties"]["$ai_total_tokens"] == 150
+
+    def test_error_type_categorization_model_behavior(
+        self, processor, mock_client, mock_span
+    ):
+        """Test that ModelBehaviorError is categorized correctly."""
+        span_data = GenerationSpanData(model="gpt-4o")
+        mock_span.span_data = span_data
+        mock_span.error = {
+            "message": "ModelBehaviorError: Invalid JSON output",
+            "type": "ModelBehaviorError",
+        }
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+        assert call_kwargs["properties"]["$ai_error_type"] == "model_behavior_error"
+
+    def test_error_type_categorization_user_error(
+        self, processor, mock_client, mock_span
+    ):
+        """Test that UserError is categorized correctly."""
+        span_data = GenerationSpanData(model="gpt-4o")
+        mock_span.span_data = span_data
+        mock_span.error = {"message": "UserError: Tool failed", "type": "UserError"}
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+        assert call_kwargs["properties"]["$ai_error_type"] == "user_error"
+
+    def test_error_type_categorization_input_guardrail(
+        self, processor, mock_client, mock_span
+    ):
+        """Test that InputGuardrailTripwireTriggered is categorized correctly."""
+        span_data = GenerationSpanData(model="gpt-4o")
+        mock_span.span_data = span_data
+        mock_span.error = {
+            "message": "InputGuardrailTripwireTriggered: Content blocked"
+        }
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+        assert (
+            call_kwargs["properties"]["$ai_error_type"] == "input_guardrail_triggered"
+        )
+
+    def test_error_type_categorization_output_guardrail(
+        self, processor, mock_client, mock_span
+    ):
+        """Test that OutputGuardrailTripwireTriggered is categorized correctly."""
+        span_data = GenerationSpanData(model="gpt-4o")
+        mock_span.span_data = span_data
+        mock_span.error = {
+            "message": "OutputGuardrailTripwireTriggered: Response blocked"
+        }
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+        assert (
+            call_kwargs["properties"]["$ai_error_type"] == "output_guardrail_triggered"
+        )
+
+    def test_error_type_categorization_max_turns(
+        self, processor, mock_client, mock_span
+    ):
+        """Test that MaxTurnsExceeded is categorized correctly."""
+        span_data = GenerationSpanData(model="gpt-4o")
+        mock_span.span_data = span_data
+        mock_span.error = {"message": "MaxTurnsExceeded: Agent exceeded maximum turns"}
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+        assert call_kwargs["properties"]["$ai_error_type"] == "max_turns_exceeded"
+
+    def test_error_type_categorization_unknown(self, processor, mock_client, mock_span):
+        """Test that unknown errors are categorized as unknown."""
+        span_data = GenerationSpanData(model="gpt-4o")
+        mock_span.span_data = span_data
+        mock_span.error = {"message": "Some random error occurred"}
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+        assert call_kwargs["properties"]["$ai_error_type"] == "unknown"
+
+    def test_response_span_with_output_and_total_tokens(
+        self, processor, mock_client, mock_span
+    ):
+        """Test ResponseSpanData includes output choices and total tokens."""
+        # Create a mock response object
+        mock_response = MagicMock()
+        mock_response.id = "resp_123"
+        mock_response.model = "gpt-4o"
+        mock_response.output = [{"type": "message", "content": "Hello!"}]
+        mock_response.usage = MagicMock()
+        mock_response.usage.input_tokens = 25
+        mock_response.usage.output_tokens = 10
+
+        span_data = ResponseSpanData(
+            response=mock_response,
+            input="Hello, world!",
+        )
+        mock_span.span_data = span_data
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+
+        assert call_kwargs["event"] == "$ai_generation"
+        assert call_kwargs["properties"]["$ai_total_tokens"] == 35
+        assert call_kwargs["properties"]["$ai_output_choices"] == [
+            {"type": "message", "content": "Hello!"}
+        ]
+        assert call_kwargs["properties"]["$ai_response_id"] == "resp_123"
+
+    def test_speech_span_with_pass_through_properties(
+        self, processor, mock_client, mock_span
+    ):
+        """Test SpeechSpanData includes pass-through properties."""
+        span_data = SpeechSpanData(
+            input="Hello, how can I help you?",
+            output="base64_audio_data",
+            output_format="pcm",
+            model="tts-1",
+            model_config={"voice": "alloy", "speed": 1.0},
+            first_content_at="2024-01-01T00:00:00.500Z",
+        )
+        mock_span.span_data = span_data
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+
+        assert call_kwargs["event"] == "$ai_span"
+        assert call_kwargs["properties"]["$ai_span_type"] == "speech"
+        assert call_kwargs["properties"]["$ai_model"] == "tts-1"
+        # Pass-through properties (no $ai_ prefix)
+        assert (
+            call_kwargs["properties"]["first_content_at"] == "2024-01-01T00:00:00.500Z"
+        )
+        assert call_kwargs["properties"]["audio_output_format"] == "pcm"
+        assert call_kwargs["properties"]["model_config"] == {
+            "voice": "alloy",
+            "speed": 1.0,
+        }
+        # Text input should be captured
+        assert call_kwargs["properties"]["$ai_input"] == "Hello, how can I help you?"
+
+    def test_transcription_span_with_pass_through_properties(
+        self, processor, mock_client, mock_span
+    ):
+        """Test TranscriptionSpanData includes pass-through properties."""
+        span_data = TranscriptionSpanData(
+            input="base64_audio_data",
+            input_format="pcm",
+            output="This is the transcribed text.",
+            model="whisper-1",
+            model_config={"language": "en"},
+        )
+        mock_span.span_data = span_data
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+
+        assert call_kwargs["event"] == "$ai_span"
+        assert call_kwargs["properties"]["$ai_span_type"] == "transcription"
+        assert call_kwargs["properties"]["$ai_model"] == "whisper-1"
+        # Pass-through properties (no $ai_ prefix)
+        assert call_kwargs["properties"]["audio_input_format"] == "pcm"
+        assert call_kwargs["properties"]["model_config"] == {"language": "en"}
+        # Transcription output should be captured
+        assert (
+            call_kwargs["properties"]["$ai_output_state"]
+            == "This is the transcribed text."
+        )
+
+    def test_latency_calculation(self, processor, mock_client, mock_span):
+        """Test that latency is calculated correctly."""
+        span_data = GenerationSpanData(model="gpt-4o")
+        mock_span.span_data = span_data
+
+        with patch("time.time") as mock_time:
+            mock_time.return_value = 1000.0
+            processor.on_span_start(mock_span)
+
+            mock_time.return_value = 1001.5  # 1.5 seconds later
+            processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+        assert call_kwargs["properties"]["$ai_latency"] == pytest.approx(1.5, rel=0.01)
+
+    def test_groups_included_in_events(self, mock_client, mock_trace, mock_span):
+        """Test that groups are included in captured events."""
+        processor = InsightsTracingProcessor(
+            client=mock_client,
+            distinct_id="test-user",
+            groups={"company": "acme", "team": "engineering"},
+        )
+
+        processor.on_trace_start(mock_trace)
+        processor.on_trace_end(mock_trace)
+
+        call_kwargs = mock_client.capture.call_args[1]
+        assert call_kwargs["groups"] == {"company": "acme", "team": "engineering"}
+
+    def test_additional_properties_included(self, mock_client, mock_trace):
+        """Test that additional properties are included in events."""
+        processor = InsightsTracingProcessor(
+            client=mock_client,
+            distinct_id="test-user",
+            properties={"environment": "production", "version": "1.0"},
+        )
+
+        processor.on_trace_start(mock_trace)
+        processor.on_trace_end(mock_trace)
+
+        call_kwargs = mock_client.capture.call_args[1]
+        assert call_kwargs["properties"]["environment"] == "production"
+        assert call_kwargs["properties"]["version"] == "1.0"
+
+    def test_shutdown_clears_state(self, processor):
+        """Test that shutdown clears internal state."""
+        processor._span_start_times["span_1"] = 1000.0
+        processor._trace_metadata["trace_1"] = {"name": "test"}
+
+        processor.shutdown()
+
+        assert len(processor._span_start_times) == 0
+        assert len(processor._trace_metadata) == 0
+
+    def test_force_flush_calls_client_flush(self, processor, mock_client):
+        """Test that force_flush calls client.flush()."""
+        processor.force_flush()
+        mock_client.flush.assert_called_once()
+
+    def test_generation_span_with_no_usage(self, processor, mock_client, mock_span):
+        """Test GenerationSpanData with no usage data defaults to zero tokens."""
+        span_data = GenerationSpanData(model="gpt-4o")
+        mock_span.span_data = span_data
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+        assert call_kwargs["properties"]["$ai_input_tokens"] == 0
+        assert call_kwargs["properties"]["$ai_output_tokens"] == 0
+        assert call_kwargs["properties"]["$ai_total_tokens"] == 0
+
+    def test_generation_span_with_partial_usage(
+        self, processor, mock_client, mock_span
+    ):
+        """Test GenerationSpanData with only input_tokens present."""
+        span_data = GenerationSpanData(
+            model="gpt-4o",
+            usage={"input_tokens": 42},
+        )
+        mock_span.span_data = span_data
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+        assert call_kwargs["properties"]["$ai_input_tokens"] == 42
+        assert call_kwargs["properties"]["$ai_output_tokens"] == 0
+        assert call_kwargs["properties"]["$ai_total_tokens"] == 42
+
+    def test_error_type_categorization_by_type_field_only(
+        self, processor, mock_client, mock_span
+    ):
+        """Test error categorization works when only the type field matches."""
+        span_data = GenerationSpanData(model="gpt-4o")
+        mock_span.span_data = span_data
+        mock_span.error = {
+            "message": "Something went wrong",
+            "type": "ModelBehaviorError",
+        }
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+        assert call_kwargs["properties"]["$ai_error_type"] == "model_behavior_error"
+
+    def test_distinct_id_resolved_from_trace_for_spans(
+        self, mock_client, mock_trace, mock_span
+    ):
+        """Test that spans use the distinct_id resolved at trace start."""
+
+        def resolver(trace):
+            return f"user-{trace.name}"
+
+        processor = InsightsTracingProcessor(
+            client=mock_client,
+            distinct_id=resolver,
+        )
+
+        # Start trace - this resolves and stores distinct_id
+        processor.on_trace_start(mock_trace)
+        mock_client.capture.reset_mock()
+
+        # End a span - should use the stored distinct_id from trace
+        span_data = GenerationSpanData(model="gpt-4o")
+        mock_span.span_data = span_data
+
+        processor.on_span_start(mock_span)
+        processor.on_span_end(mock_span)
+
+        call_kwargs = mock_client.capture.call_args[1]
+        assert call_kwargs["distinct_id"] == "user-Test Workflow"
+
+    def test_eviction_of_stale_entries(self, mock_client):
+        """Test that stale entries are evicted when max is exceeded."""
+        processor = InsightsTracingProcessor(
+            client=mock_client,
+            distinct_id="test-user",
+        )
+        processor._max_tracked_entries = 10
+
+        # Fill beyond max
+        for i in range(15):
+            processor._span_start_times[f"span_{i}"] = float(i)
+            processor._trace_metadata[f"trace_{i}"] = {"name": f"trace_{i}"}
+
+        processor._evict_stale_entries()
+
+        # Should have evicted half
+        assert len(processor._span_start_times) <= 10
+        assert len(processor._trace_metadata) <= 10
+
+
+class TestInstrumentHelper:
+    """Tests for the instrument() convenience function."""
+
+    def test_instrument_registers_processor(self, mock_client):
+        """Test that instrument() registers a processor."""
+        with patch("agents.tracing.add_trace_processor") as mock_add:
+            processor = instrument(
+                client=mock_client,
+                distinct_id="test-user",
+            )
+
+            mock_add.assert_called_once_with(processor)
+            assert isinstance(processor, InsightsTracingProcessor)
+
+    def test_instrument_with_privacy_mode(self, mock_client):
+        """Test instrument() respects privacy_mode."""
+        with patch("agents.tracing.add_trace_processor"):
+            processor = instrument(
+                client=mock_client,
+                privacy_mode=True,
+            )
+
+            assert processor._privacy_mode is True
+
+    def test_instrument_with_groups_and_properties(self, mock_client):
+        """Test instrument() accepts groups and properties."""
+        with patch("agents.tracing.add_trace_processor"):
+            processor = instrument(
+                client=mock_client,
+                groups={"company": "acme"},
+                properties={"env": "test"},
+            )
+
+            assert processor._groups == {"company": "acme"}
+            assert processor._properties == {"env": "test"}

hanzo_insights/test/ai/test_prompts.py

@@ -0,0 +1,764 @@
+import unittest
+from unittest.mock import MagicMock, patch
+
+from hanzo_insights.ai.prompts import Prompts
+
+
+class MockResponse:
+    """Mock HTTP response for testing."""
+
+    def __init__(self, json_data=None, status_code=200, ok=True):
+        self._json_data = json_data
+        self.status_code = status_code
+        self.ok = ok
+
+    def json(self):
+        if self._json_data is None:
+            raise ValueError("No JSON data")
+        return self._json_data
+
+
+class TestPrompts(unittest.TestCase):
+    """Tests for the Prompts class."""
+
+    mock_prompt_response = {
+        "id": 1,
+        "name": "test-prompt",
+        "prompt": "Hello, {{name}}! You are a helpful assistant for {{company}}.",
+        "version": 1,
+        "created_by": "user@example.com",
+        "created_at": "2024-01-01T00:00:00Z",
+        "updated_at": "2024-01-01T00:00:00Z",
+        "deleted": False,
+    }
+
+    def create_mock_client(
+        self,
+        personal_api_key="phx_test_key",
+        project_api_key="phc_test_key",
+        host="https://us.insights.hanzo.ai",
+    ):
+        """Create a mock Insights client."""
+        mock = MagicMock()
+        mock.personal_api_key = personal_api_key
+        mock.api_key = project_api_key
+        mock.raw_host = host
+        return mock
+
+
+class TestPromptsGet(TestPrompts):
+    """Tests for the Prompts.get() method."""
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    def test_successfully_fetch_a_prompt(self, mock_get_session):
+        """Should successfully fetch a prompt."""
+        mock_get = mock_get_session.return_value.get
+        mock_get.return_value = MockResponse(json_data=self.mock_prompt_response)
+
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        result = prompts.get("test-prompt")
+
+        self.assertEqual(result, self.mock_prompt_response["prompt"])
+        mock_get.assert_called_once()
+        call_args = mock_get.call_args
+        self.assertEqual(
+            call_args[0][0],
+            "https://us.insights.hanzo.ai/api/environments/@current/llm_prompts/name/test-prompt/?token=phc_test_key",
+        )
+        self.assertIn("Authorization", call_args[1]["headers"])
+        self.assertEqual(
+            call_args[1]["headers"]["Authorization"], "Bearer phx_test_key"
+        )
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    def test_successfully_fetch_a_specific_prompt_version(self, mock_get_session):
+        """Should successfully fetch a specific prompt version."""
+        mock_get = mock_get_session.return_value.get
+        versioned_prompt_response = {
+            **self.mock_prompt_response,
+            "prompt": "Prompt version 1",
+            "version": 1,
+        }
+        mock_get.return_value = MockResponse(json_data=versioned_prompt_response)
+
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        result = prompts.get("test-prompt", version=1)
+
+        self.assertEqual(result, versioned_prompt_response["prompt"])
+        mock_get.assert_called_once()
+        call_args = mock_get.call_args
+        self.assertEqual(
+            call_args[0][0],
+            "https://us.insights.hanzo.ai/api/environments/@current/llm_prompts/name/test-prompt/?token=phc_test_key&version=1",
+        )
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    @patch("hanzo_insights.ai.prompts.time.time")
+    def test_return_cached_prompt_when_fresh(self, mock_time, mock_get_session):
+        """Should return cached prompt when fresh (no API call)."""
+        mock_get = mock_get_session.return_value.get
+        mock_get.return_value = MockResponse(json_data=self.mock_prompt_response)
+        mock_time.return_value = 1000.0
+
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        # First call - fetches from API
+        result1 = prompts.get("test-prompt", cache_ttl_seconds=300)
+        self.assertEqual(result1, self.mock_prompt_response["prompt"])
+        self.assertEqual(mock_get.call_count, 1)
+
+        # Advance time by 60 seconds (still within TTL)
+        mock_time.return_value = 1060.0
+
+        # Second call - should use cache
+        result2 = prompts.get("test-prompt", cache_ttl_seconds=300)
+        self.assertEqual(result2, self.mock_prompt_response["prompt"])
+        self.assertEqual(mock_get.call_count, 1)  # No additional fetch
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    def test_cache_latest_and_versioned_prompts_separately(self, mock_get_session):
+        """Should cache latest and historical prompt versions separately."""
+        mock_get = mock_get_session.return_value.get
+        latest_prompt_response = {
+            **self.mock_prompt_response,
+            "prompt": "Latest prompt",
+            "version": 2,
+        }
+        versioned_prompt_response = {
+            **self.mock_prompt_response,
+            "prompt": "Prompt version 1",
+            "version": 1,
+        }
+
+        mock_get.side_effect = [
+            MockResponse(json_data=latest_prompt_response),
+            MockResponse(json_data=versioned_prompt_response),
+        ]
+
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        self.assertEqual(prompts.get("test-prompt"), latest_prompt_response["prompt"])
+        self.assertEqual(
+            prompts.get("test-prompt", version=1),
+            versioned_prompt_response["prompt"],
+        )
+        self.assertEqual(prompts.get("test-prompt"), latest_prompt_response["prompt"])
+        self.assertEqual(
+            prompts.get("test-prompt", version=1),
+            versioned_prompt_response["prompt"],
+        )
+        self.assertEqual(mock_get.call_count, 2)
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    @patch("hanzo_insights.ai.prompts.time.time")
+    def test_refetch_when_cache_is_stale(self, mock_time, mock_get_session):
+        """Should refetch when cache is stale."""
+        mock_get = mock_get_session.return_value.get
+        updated_prompt_response = {
+            **self.mock_prompt_response,
+            "prompt": "Updated prompt: Hello, {{name}}!",
+        }
+
+        mock_get.side_effect = [
+            MockResponse(json_data=self.mock_prompt_response),
+            MockResponse(json_data=updated_prompt_response),
+        ]
+        mock_time.return_value = 1000.0
+
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        # First call - fetches from API
+        result1 = prompts.get("test-prompt", cache_ttl_seconds=60)
+        self.assertEqual(result1, self.mock_prompt_response["prompt"])
+        self.assertEqual(mock_get.call_count, 1)
+
+        # Advance time past TTL
+        mock_time.return_value = 1061.0
+
+        # Second call - should refetch
+        result2 = prompts.get("test-prompt", cache_ttl_seconds=60)
+        self.assertEqual(result2, updated_prompt_response["prompt"])
+        self.assertEqual(mock_get.call_count, 2)
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    @patch("hanzo_insights.ai.prompts.time.time")
+    @patch("hanzo_insights.ai.prompts.log")
+    def test_use_stale_cache_on_fetch_failure_with_warning(
+        self, mock_log, mock_time, mock_get_session
+    ):
+        """Should use stale cache on fetch failure with warning."""
+        mock_get = mock_get_session.return_value.get
+        mock_get.side_effect = [
+            MockResponse(json_data=self.mock_prompt_response),
+            Exception("Network error"),
+        ]
+        mock_time.return_value = 1000.0
+
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        # First call - populates cache
+        result1 = prompts.get("test-prompt", cache_ttl_seconds=60)
+        self.assertEqual(result1, self.mock_prompt_response["prompt"])
+
+        # Advance time past TTL
+        mock_time.return_value = 1061.0
+
+        # Second call - should use stale cache
+        result2 = prompts.get("test-prompt", cache_ttl_seconds=60)
+        self.assertEqual(result2, self.mock_prompt_response["prompt"])
+
+        # Check warning was logged
+        mock_log.warning.assert_called()
+        warning_call = mock_log.warning.call_args
+        self.assertIn("using stale cache", warning_call[0][0])
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    @patch("hanzo_insights.ai.prompts.log")
+    def test_use_fallback_when_no_cache_and_fetch_fails_with_warning(
+        self, mock_log, mock_get_session
+    ):
+        """Should use fallback when no cache and fetch fails with warning."""
+        mock_get = mock_get_session.return_value.get
+        mock_get.side_effect = Exception("Network error")
+
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        fallback = "Default system prompt."
+        result = prompts.get("test-prompt", fallback=fallback)
+
+        self.assertEqual(result, fallback)
+
+        # Check warning was logged
+        mock_log.warning.assert_called()
+        warning_call = mock_log.warning.call_args
+        self.assertIn("using fallback", warning_call[0][0])
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    def test_throw_when_no_cache_no_fallback_and_fetch_fails(self, mock_get_session):
+        """Should throw when no cache, no fallback, and fetch fails."""
+        mock_get = mock_get_session.return_value.get
+        mock_get.side_effect = Exception("Network error")
+
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        with self.assertRaises(Exception) as context:
+            prompts.get("test-prompt")
+
+        self.assertIn("Network error", str(context.exception))
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    def test_handle_404_response(self, mock_get_session):
+        """Should handle 404 response."""
+        mock_get = mock_get_session.return_value.get
+        mock_get.return_value = MockResponse(status_code=404, ok=False)
+
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        with self.assertRaises(Exception) as context:
+            prompts.get("nonexistent-prompt")
+
+        self.assertIn('Prompt "nonexistent-prompt" not found', str(context.exception))
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    def test_handle_404_response_for_specific_prompt_version(self, mock_get_session):
+        """Should handle 404 response for a specific prompt version."""
+        mock_get = mock_get_session.return_value.get
+        mock_get.return_value = MockResponse(status_code=404, ok=False)
+
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        with self.assertRaises(Exception) as context:
+            prompts.get("nonexistent-prompt", version=3)
+
+        self.assertIn(
+            'Prompt "nonexistent-prompt" version 3 not found',
+            str(context.exception),
+        )
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    def test_handle_403_response(self, mock_get_session):
+        """Should handle 403 response."""
+        mock_get = mock_get_session.return_value.get
+        mock_get.return_value = MockResponse(status_code=403, ok=False)
+
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        with self.assertRaises(Exception) as context:
+            prompts.get("restricted-prompt")
+
+        self.assertIn(
+            'Access denied for prompt "restricted-prompt"', str(context.exception)
+        )
+
+    def test_throw_when_no_personal_api_key_configured(self):
+        """Should throw when no personal_api_key is configured."""
+        client = self.create_mock_client(personal_api_key=None)
+        prompts = Prompts(client)
+
+        with self.assertRaises(Exception) as context:
+            prompts.get("test-prompt")
+
+        self.assertIn(
+            "personal_api_key is required to fetch prompts", str(context.exception)
+        )
+
+    def test_throw_when_no_project_api_key_configured(self):
+        """Should throw when no project_api_key is configured."""
+        client = self.create_mock_client(project_api_key=None)
+        prompts = Prompts(client)
+
+        with self.assertRaises(Exception) as context:
+            prompts.get("test-prompt")
+
+        self.assertIn(
+            "project_api_key is required to fetch prompts", str(context.exception)
+        )
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    def test_throw_when_api_returns_invalid_response_format(self, mock_get_session):
+        """Should throw when API returns invalid response format."""
+        mock_get = mock_get_session.return_value.get
+        mock_get.return_value = MockResponse(json_data={"invalid": "response"})
+
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        with self.assertRaises(Exception) as context:
+            prompts.get("test-prompt")
+
+        self.assertIn("Invalid response format", str(context.exception))
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    def test_use_custom_host_from_insights_options(self, mock_get_session):
+        """Should use custom host from Insights options."""
+        mock_get = mock_get_session.return_value.get
+        mock_get.return_value = MockResponse(json_data=self.mock_prompt_response)
+
+        client = self.create_mock_client(host="https://eu.insights.hanzo.ai")
+        prompts = Prompts(client)
+
+        prompts.get("test-prompt")
+
+        call_args = mock_get.call_args
+        self.assertTrue(
+            call_args[0][0].startswith(
+                "https://eu.insights.hanzo.ai/api/environments/@current/llm_prompts/name/test-prompt/?token=phc_test_key"
+            ),
+            f"Expected URL to start with 'https://eu.insights.hanzo.ai/api/environments/@current/llm_prompts/name/test-prompt/?token=phc_test_key', got {call_args[0][0]}",
+        )
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    @patch("hanzo_insights.ai.prompts.time.time")
+    def test_use_default_cache_ttl_5_minutes(self, mock_time, mock_get_session):
+        """Should use default cache TTL (5 minutes) when not specified."""
+        mock_get = mock_get_session.return_value.get
+        mock_get.return_value = MockResponse(json_data=self.mock_prompt_response)
+        mock_time.return_value = 1000.0
+
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        # First call
+        prompts.get("test-prompt")
+        self.assertEqual(mock_get.call_count, 1)
+
+        # Advance time by 4 minutes (within default 5-minute TTL)
+        mock_time.return_value = 1000.0 + (4 * 60)
+
+        # Second call - should use cache
+        prompts.get("test-prompt")
+        self.assertEqual(mock_get.call_count, 1)
+
+        # Advance time past 5-minute TTL
+        mock_time.return_value = 1000.0 + (6 * 60)
+
+        # Third call - should refetch
+        prompts.get("test-prompt")
+        self.assertEqual(mock_get.call_count, 2)
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    @patch("hanzo_insights.ai.prompts.time.time")
+    def test_use_custom_default_cache_ttl_from_constructor(
+        self, mock_time, mock_get_session
+    ):
+        """Should use custom default cache TTL from constructor."""
+        mock_get = mock_get_session.return_value.get
+        mock_get.return_value = MockResponse(json_data=self.mock_prompt_response)
+        mock_time.return_value = 1000.0
+
+        client = self.create_mock_client()
+        prompts = Prompts(client, default_cache_ttl_seconds=60)
+
+        # First call
+        prompts.get("test-prompt")
+        self.assertEqual(mock_get.call_count, 1)
+
+        # Advance time past custom TTL
+        mock_time.return_value = 1061.0
+
+        # Second call - should refetch
+        prompts.get("test-prompt")
+        self.assertEqual(mock_get.call_count, 2)
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    def test_url_encode_prompt_names_with_special_characters(self, mock_get_session):
+        """Should URL-encode prompt names with special characters."""
+        mock_get = mock_get_session.return_value.get
+        mock_get.return_value = MockResponse(json_data=self.mock_prompt_response)
+
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        prompts.get("prompt with spaces/and/slashes")
+
+        call_args = mock_get.call_args
+        self.assertEqual(
+            call_args[0][0],
+            "https://us.insights.hanzo.ai/api/environments/@current/llm_prompts/name/prompt%20with%20spaces%2Fand%2Fslashes/?token=phc_test_key",
+        )
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    def test_work_with_direct_options_no_insights_client(self, mock_get_session):
+        """Should work with direct options (no Insights client)."""
+        mock_get = mock_get_session.return_value.get
+        mock_get.return_value = MockResponse(json_data=self.mock_prompt_response)
+
+        prompts = Prompts(
+            personal_api_key="phx_direct_key", project_api_key="phc_direct_key"
+        )
+
+        result = prompts.get("test-prompt")
+
+        self.assertEqual(result, self.mock_prompt_response["prompt"])
+        call_args = mock_get.call_args
+        self.assertEqual(
+            call_args[0][0],
+            "https://us.insights.hanzo.ai/api/environments/@current/llm_prompts/name/test-prompt/?token=phc_direct_key",
+        )
+        self.assertEqual(
+            call_args[1]["headers"]["Authorization"], "Bearer phx_direct_key"
+        )
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    def test_use_custom_host_from_direct_options(self, mock_get_session):
+        """Should use custom host from direct options."""
+        mock_get = mock_get_session.return_value.get
+        mock_get.return_value = MockResponse(json_data=self.mock_prompt_response)
+
+        prompts = Prompts(
+            personal_api_key="phx_direct_key",
+            project_api_key="phc_direct_key",
+            host="https://eu.insights.hanzo.ai",
+        )
+
+        prompts.get("test-prompt")
+
+        call_args = mock_get.call_args
+        self.assertEqual(
+            call_args[0][0],
+            "https://eu.insights.hanzo.ai/api/environments/@current/llm_prompts/name/test-prompt/?token=phc_direct_key",
+        )
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    @patch("hanzo_insights.ai.prompts.time.time")
+    def test_use_custom_default_cache_ttl_from_direct_options(
+        self, mock_time, mock_get_session
+    ):
+        """Should use custom default cache TTL from direct options."""
+        mock_get = mock_get_session.return_value.get
+        mock_get.return_value = MockResponse(json_data=self.mock_prompt_response)
+        mock_time.return_value = 1000.0
+
+        prompts = Prompts(
+            personal_api_key="phx_direct_key",
+            project_api_key="phc_direct_key",
+            default_cache_ttl_seconds=60,
+        )
+
+        # First call
+        prompts.get("test-prompt")
+        self.assertEqual(mock_get.call_count, 1)
+
+        # Advance time past custom TTL
+        mock_time.return_value = 1061.0
+
+        # Second call - should refetch
+        prompts.get("test-prompt")
+        self.assertEqual(mock_get.call_count, 2)
+
+
+class TestPromptsCompile(TestPrompts):
+    """Tests for the Prompts.compile() method."""
+
+    def test_replace_a_single_variable(self):
+        """Should replace a single variable."""
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        result = prompts.compile("Hello, {{name}}!", {"name": "World"})
+
+        self.assertEqual(result, "Hello, World!")
+
+    def test_replace_multiple_variables(self):
+        """Should replace multiple variables."""
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        result = prompts.compile(
+            "Hello, {{name}}! Welcome to {{company}}. Your tier is {{tier}}.",
+            {"name": "John", "company": "Acme Corp", "tier": "premium"},
+        )
+
+        self.assertEqual(
+            result, "Hello, John! Welcome to Acme Corp. Your tier is premium."
+        )
+
+    def test_handle_numbers(self):
+        """Should handle numbers."""
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        result = prompts.compile("You have {{count}} items.", {"count": 42})
+
+        self.assertEqual(result, "You have 42 items.")
+
+    def test_handle_booleans(self):
+        """Should handle booleans."""
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        result = prompts.compile("Feature enabled: {{enabled}}", {"enabled": True})
+
+        self.assertEqual(result, "Feature enabled: True")
+
+    def test_leave_unmatched_variables_unchanged(self):
+        """Should leave unmatched variables unchanged."""
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        result = prompts.compile(
+            "Hello, {{name}}! Your {{unknown}} is ready.", {"name": "World"}
+        )
+
+        self.assertEqual(result, "Hello, World! Your {{unknown}} is ready.")
+
+    def test_handle_prompts_with_no_variables(self):
+        """Should handle prompts with no variables."""
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        result = prompts.compile("You are a helpful assistant.", {})
+
+        self.assertEqual(result, "You are a helpful assistant.")
+
+    def test_handle_empty_variables_dict(self):
+        """Should handle empty variables dict."""
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        result = prompts.compile("Hello, {{name}}!", {})
+
+        self.assertEqual(result, "Hello, {{name}}!")
+
+    def test_handle_multiple_occurrences_of_same_variable(self):
+        """Should handle multiple occurrences of the same variable."""
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        result = prompts.compile(
+            "Hello, {{name}}! Goodbye, {{name}}!", {"name": "World"}
+        )
+
+        self.assertEqual(result, "Hello, World! Goodbye, World!")
+
+    def test_work_with_direct_options_initialization(self):
+        """Should work with direct options initialization."""
+        prompts = Prompts(
+            personal_api_key="phx_test_key", project_api_key="phc_test_key"
+        )
+
+        result = prompts.compile("Hello, {{name}}!", {"name": "World"})
+
+        self.assertEqual(result, "Hello, World!")
+
+    def test_handle_variables_with_hyphens(self):
+        """Should handle variables with hyphens."""
+        prompts = Prompts(
+            personal_api_key="phx_test_key", project_api_key="phc_test_key"
+        )
+
+        result = prompts.compile("User ID: {{user-id}}", {"user-id": "12345"})
+
+        self.assertEqual(result, "User ID: 12345")
+
+    def test_handle_variables_with_dots(self):
+        """Should handle variables with dots."""
+        prompts = Prompts(
+            personal_api_key="phx_test_key", project_api_key="phc_test_key"
+        )
+
+        result = prompts.compile("Company: {{company.name}}", {"company.name": "Acme"})
+
+        self.assertEqual(result, "Company: Acme")
+
+
+class TestPromptsClearCache(TestPrompts):
+    """Tests for the Prompts.clear_cache() method."""
+
+    def test_clear_cache_with_version_and_no_name_raises_value_error(self):
+        """Should enforce that versioned cache clearing requires a prompt name."""
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        with self.assertRaises(ValueError) as context:
+            prompts.clear_cache(version=1)
+
+        self.assertIn("requires 'name'", str(context.exception))
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    def test_clear_a_specific_prompt_from_cache(self, mock_get_session):
+        """Should clear a specific prompt from cache."""
+        mock_get = mock_get_session.return_value.get
+        other_prompt_response = {**self.mock_prompt_response, "name": "other-prompt"}
+
+        mock_get.side_effect = [
+            MockResponse(json_data=self.mock_prompt_response),
+            MockResponse(json_data=other_prompt_response),
+            MockResponse(json_data=self.mock_prompt_response),
+        ]
+
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        # Populate cache with two prompts
+        prompts.get("test-prompt")
+        prompts.get("other-prompt")
+        self.assertEqual(mock_get.call_count, 2)
+
+        # Clear only test-prompt
+        prompts.clear_cache("test-prompt")
+
+        # test-prompt should be refetched
+        prompts.get("test-prompt")
+        self.assertEqual(mock_get.call_count, 3)
+
+        # other-prompt should still be cached
+        prompts.get("other-prompt")
+        self.assertEqual(mock_get.call_count, 3)
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    def test_clear_a_specific_prompt_version_from_cache(self, mock_get_session):
+        """Should clear only the requested prompt version from cache."""
+        mock_get = mock_get_session.return_value.get
+        latest_prompt_response = {
+            **self.mock_prompt_response,
+            "prompt": "Latest prompt",
+            "version": 2,
+        }
+        versioned_prompt_response = {
+            **self.mock_prompt_response,
+            "prompt": "Prompt version 1",
+            "version": 1,
+        }
+
+        mock_get.side_effect = [
+            MockResponse(json_data=latest_prompt_response),
+            MockResponse(json_data=versioned_prompt_response),
+            MockResponse(json_data=versioned_prompt_response),
+        ]
+
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        prompts.get("test-prompt")
+        prompts.get("test-prompt", version=1)
+        self.assertEqual(mock_get.call_count, 2)
+
+        prompts.clear_cache("test-prompt", version=1)
+
+        prompts.get("test-prompt")
+        self.assertEqual(mock_get.call_count, 2)
+
+        prompts.get("test-prompt", version=1)
+        self.assertEqual(mock_get.call_count, 3)
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    def test_clear_a_prompt_name_clears_all_cached_versions(self, mock_get_session):
+        """Should clear latest and versioned cache entries for the same prompt name."""
+        mock_get = mock_get_session.return_value.get
+        latest_prompt_response = {
+            **self.mock_prompt_response,
+            "prompt": "Latest prompt",
+            "version": 2,
+        }
+        versioned_prompt_response = {
+            **self.mock_prompt_response,
+            "prompt": "Prompt version 1",
+            "version": 1,
+        }
+
+        mock_get.side_effect = [
+            MockResponse(json_data=latest_prompt_response),
+            MockResponse(json_data=versioned_prompt_response),
+            MockResponse(json_data=latest_prompt_response),
+            MockResponse(json_data=versioned_prompt_response),
+        ]
+
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        prompts.get("test-prompt")
+        prompts.get("test-prompt", version=1)
+        self.assertEqual(mock_get.call_count, 2)
+
+        prompts.clear_cache("test-prompt")
+
+        prompts.get("test-prompt")
+        prompts.get("test-prompt", version=1)
+        self.assertEqual(mock_get.call_count, 4)
+
+    @patch("hanzo_insights.ai.prompts._get_session")
+    def test_clear_all_prompts_from_cache(self, mock_get_session):
+        """Should clear all prompts from cache when no name is provided."""
+        mock_get = mock_get_session.return_value.get
+        other_prompt_response = {**self.mock_prompt_response, "name": "other-prompt"}
+
+        mock_get.side_effect = [
+            MockResponse(json_data=self.mock_prompt_response),
+            MockResponse(json_data=other_prompt_response),
+            MockResponse(json_data=self.mock_prompt_response),
+            MockResponse(json_data=other_prompt_response),
+        ]
+
+        client = self.create_mock_client()
+        prompts = Prompts(client)
+
+        # Populate cache with two prompts
+        prompts.get("test-prompt")
+        prompts.get("other-prompt")
+        self.assertEqual(mock_get.call_count, 2)
+
+        # Clear all cache
+        prompts.clear_cache()
+
+        # Both prompts should be refetched
+        prompts.get("test-prompt")
+        prompts.get("other-prompt")
+        self.assertEqual(mock_get.call_count, 4)
+
+
+if __name__ == "__main__":
+    unittest.main()

hanzo_insights/test/ai/test_sanitization.py

@@ -1,7 +1,7 @@
 import os
 import unittest
 
-from posthog.ai.sanitization import (
+from hanzo_insights.ai.sanitization import (
     redact_base64_data_url,
     sanitize_openai,
     sanitize_openai_response,
@@ -69,6 +69,25 @@ class TestSanitization(unittest.TestCase):
         )
         self.assertEqual(result[0]["content"][1]["image_url"]["detail"], "high")
 
+    def test_sanitize_openai_input_image(self):
+        input_data = [
+            {
+                "role": "user",
+                "content": [
+                    {
+                        "type": "input_image",
+                        "image_url": self.sample_base64_image,
+                    }
+                ],
+            }
+        ]
+
+        result = sanitize_openai(input_data)
+
+        self.assertEqual(
+            result[0]["content"][0]["image_url"], REDACTED_IMAGE_PLACEHOLDER
+        )
+
     def test_sanitize_openai_preserves_regular_urls(self):
         input_data = [
             {

hanzo_insights/test/ai/test_system_prompts.py

@@ -11,7 +11,10 @@ regardless of how they're passed to the providers:
 
 import time
 import unittest
-from unittest.mock import patch, MagicMock
+from unittest.mock import MagicMock, patch
+
+from hanzo_insights.client import Client
+from hanzo_insights.test.test_utils import FAKE_TEST_API_KEY
 
 
 class TestSystemPromptCapture(unittest.TestCase):
@@ -23,8 +26,9 @@ class TestSystemPromptCapture(unittest.TestCase):
         self.test_user_message = "Hello, how are you?"
         self.test_response = "I'm doing well, thank you!"
 
-        # Create mock PostHog client
-        self.client = MagicMock()
+        # Create mock Insights client
+        self.client = Client(FAKE_TEST_API_KEY)
+        self.client._enqueue = MagicMock()
         self.client.privacy_mode = False
 
     def _assert_system_prompt_captured(self, captured_input):
@@ -53,10 +57,11 @@ class TestSystemPromptCapture(unittest.TestCase):
     def test_openai_messages_array_system_prompt(self):
         """Test OpenAI with system prompt in messages array."""
         try:
-            from posthog.ai.openai import OpenAI
             from openai.types.chat import ChatCompletion, ChatCompletionMessage
             from openai.types.chat.chat_completion import Choice
             from openai.types.completion_usage import CompletionUsage
+
+            from hanzo_insights.ai.openai import OpenAI
         except ImportError:
             self.skipTest("OpenAI package not available")
 
@@ -83,7 +88,7 @@ class TestSystemPromptCapture(unittest.TestCase):
             "openai.resources.chat.completions.Completions.create",
             return_value=mock_response,
         ):
-            client = OpenAI(posthog_client=self.client, api_key="test")
+            client = OpenAI(insights_client=self.client, api_key="test")
 
             messages = [
                 {"role": "system", "content": self.test_system_prompt},
@@ -91,20 +96,21 @@ class TestSystemPromptCapture(unittest.TestCase):
             ]
 
             client.chat.completions.create(
-                model="gpt-4", messages=messages, posthog_distinct_id="test-user"
+                model="gpt-4", messages=messages, insights_distinct_id="test-user"
             )
 
-            self.assertEqual(len(self.client.capture.call_args_list), 1)
-            properties = self.client.capture.call_args_list[0][1]["properties"]
+            self.assertEqual(len(self.client._enqueue.call_args_list), 1)
+            properties = self.client._enqueue.call_args_list[0][0][0]["properties"]
             self._assert_system_prompt_captured(properties["$ai_input"])
 
     def test_openai_separate_system_parameter(self):
         """Test OpenAI with system prompt as separate parameter."""
         try:
-            from posthog.ai.openai import OpenAI
             from openai.types.chat import ChatCompletion, ChatCompletionMessage
             from openai.types.chat.chat_completion import Choice
             from openai.types.completion_usage import CompletionUsage
+
+            from hanzo_insights.ai.openai import OpenAI
         except ImportError:
             self.skipTest("OpenAI package not available")
 
@@ -131,7 +137,7 @@ class TestSystemPromptCapture(unittest.TestCase):
             "openai.resources.chat.completions.Completions.create",
             return_value=mock_response,
         ):
-            client = OpenAI(posthog_client=self.client, api_key="test")
+            client = OpenAI(insights_client=self.client, api_key="test")
 
             messages = [{"role": "user", "content": self.test_user_message}]
 
@@ -139,21 +145,24 @@ class TestSystemPromptCapture(unittest.TestCase):
                 model="gpt-4",
                 messages=messages,
                 system=self.test_system_prompt,
-                posthog_distinct_id="test-user",
+                insights_distinct_id="test-user",
             )
 
-            self.assertEqual(len(self.client.capture.call_args_list), 1)
-            properties = self.client.capture.call_args_list[0][1]["properties"]
+            self.assertEqual(len(self.client._enqueue.call_args_list), 1)
+            properties = self.client._enqueue.call_args_list[0][0][0]["properties"]
             self._assert_system_prompt_captured(properties["$ai_input"])
 
     def test_openai_streaming_system_parameter(self):
         """Test OpenAI streaming with system parameter."""
         try:
-            from posthog.ai.openai import OpenAI
-            from openai.types.chat.chat_completion_chunk import ChatCompletionChunk
+            from openai.types.chat.chat_completion_chunk import (
+                ChatCompletionChunk,
+                ChoiceDelta,
+            )
             from openai.types.chat.chat_completion_chunk import Choice as ChoiceChunk
-            from openai.types.chat.chat_completion_chunk import ChoiceDelta
             from openai.types.completion_usage import CompletionUsage
+
+            from hanzo_insights.ai.openai import OpenAI
         except ImportError:
             self.skipTest("OpenAI package not available")
 
@@ -192,7 +201,7 @@ class TestSystemPromptCapture(unittest.TestCase):
             "openai.resources.chat.completions.Completions.create",
             return_value=[chunk1, chunk2],
         ):
-            client = OpenAI(posthog_client=self.client, api_key="test")
+            client = OpenAI(insights_client=self.client, api_key="test")
 
             messages = [{"role": "user", "content": self.test_user_message}]
 
@@ -201,20 +210,20 @@ class TestSystemPromptCapture(unittest.TestCase):
                 messages=messages,
                 system=self.test_system_prompt,
                 stream=True,
-                posthog_distinct_id="test-user",
+                insights_distinct_id="test-user",
             )
 
             list(response_generator)  # Consume generator
 
-            self.assertEqual(len(self.client.capture.call_args_list), 1)
-            properties = self.client.capture.call_args_list[0][1]["properties"]
+            self.assertEqual(len(self.client._enqueue.call_args_list), 1)
+            properties = self.client._enqueue.call_args_list[0][0][0]["properties"]
             self._assert_system_prompt_captured(properties["$ai_input"])
 
     # Anthropic Tests
     def test_anthropic_messages_array_system_prompt(self):
         """Test Anthropic with system prompt in messages array."""
         try:
-            from posthog.ai.anthropic import Anthropic
+            from hanzo_insights.ai.anthropic import Anthropic
         except ImportError:
             self.skipTest("Anthropic package not available")
 
@@ -226,7 +235,7 @@ class TestSystemPromptCapture(unittest.TestCase):
             mock_response.usage.cache_creation_input_tokens = None
             mock_create.return_value = mock_response
 
-            client = Anthropic(posthog_client=self.client, api_key="test")
+            client = Anthropic(insights_client=self.client, api_key="test")
 
             messages = [
                 {"role": "system", "content": self.test_system_prompt},
@@ -236,17 +245,17 @@ class TestSystemPromptCapture(unittest.TestCase):
             client.messages.create(
                 model="claude-3-5-sonnet-20241022",
                 messages=messages,
-                posthog_distinct_id="test-user",
+                insights_distinct_id="test-user",
             )
 
-            self.assertEqual(len(self.client.capture.call_args_list), 1)
-            properties = self.client.capture.call_args_list[0][1]["properties"]
+            self.assertEqual(len(self.client._enqueue.call_args_list), 1)
+            properties = self.client._enqueue.call_args_list[0][0][0]["properties"]
             self._assert_system_prompt_captured(properties["$ai_input"])
 
     def test_anthropic_separate_system_parameter(self):
         """Test Anthropic with system prompt as separate parameter."""
         try:
-            from posthog.ai.anthropic import Anthropic
+            from hanzo_insights.ai.anthropic import Anthropic
         except ImportError:
             self.skipTest("Anthropic package not available")
 
@@ -258,7 +267,7 @@ class TestSystemPromptCapture(unittest.TestCase):
             mock_response.usage.cache_creation_input_tokens = None
             mock_create.return_value = mock_response
 
-            client = Anthropic(posthog_client=self.client, api_key="test")
+            client = Anthropic(insights_client=self.client, api_key="test")
 
             messages = [{"role": "user", "content": self.test_user_message}]
 
@@ -266,18 +275,18 @@ class TestSystemPromptCapture(unittest.TestCase):
                 model="claude-3-5-sonnet-20241022",
                 messages=messages,
                 system=self.test_system_prompt,
-                posthog_distinct_id="test-user",
+                insights_distinct_id="test-user",
             )
 
-            self.assertEqual(len(self.client.capture.call_args_list), 1)
-            properties = self.client.capture.call_args_list[0][1]["properties"]
+            self.assertEqual(len(self.client._enqueue.call_args_list), 1)
+            properties = self.client._enqueue.call_args_list[0][0][0]["properties"]
             self._assert_system_prompt_captured(properties["$ai_input"])
 
     # Gemini Tests
     def test_gemini_contents_array_system_prompt(self):
         """Test Gemini with system prompt in contents array."""
         try:
-            from posthog.ai.gemini import Client
+            from hanzo_insights.ai.gemini import Client
         except ImportError:
             self.skipTest("Gemini package not available")
 
@@ -297,7 +306,7 @@ class TestSystemPromptCapture(unittest.TestCase):
             mock_client_instance.models = mock_models_instance
             mock_genai_class.return_value = mock_client_instance
 
-            client = Client(posthog_client=self.client, api_key="test")
+            client = Client(insights_client=self.client, api_key="test")
 
             contents = [
                 {"role": "system", "content": self.test_system_prompt},
@@ -307,17 +316,17 @@ class TestSystemPromptCapture(unittest.TestCase):
             client.models.generate_content(
                 model="gemini-2.0-flash",
                 contents=contents,
-                posthog_distinct_id="test-user",
+                insights_distinct_id="test-user",
             )
 
-            self.assertEqual(len(self.client.capture.call_args_list), 1)
-            properties = self.client.capture.call_args_list[0][1]["properties"]
+            self.assertEqual(len(self.client._enqueue.call_args_list), 1)
+            properties = self.client._enqueue.call_args_list[0][0][0]["properties"]
             self._assert_system_prompt_captured(properties["$ai_input"])
 
     def test_gemini_system_instruction_parameter(self):
         """Test Gemini with system_instruction in config parameter."""
         try:
-            from posthog.ai.gemini import Client
+            from hanzo_insights.ai.gemini import Client
         except ImportError:
             self.skipTest("Gemini package not available")
 
@@ -337,7 +346,7 @@ class TestSystemPromptCapture(unittest.TestCase):
             mock_client_instance.models = mock_models_instance
             mock_genai_class.return_value = mock_client_instance
 
-            client = Client(posthog_client=self.client, api_key="test")
+            client = Client(insights_client=self.client, api_key="test")
 
             contents = [{"role": "user", "content": self.test_user_message}]
             config = {"system_instruction": self.test_system_prompt}
@@ -346,9 +355,9 @@ class TestSystemPromptCapture(unittest.TestCase):
                 model="gemini-2.0-flash",
                 contents=contents,
                 config=config,
-                posthog_distinct_id="test-user",
+                insights_distinct_id="test-user",
             )
 
-            self.assertEqual(len(self.client.capture.call_args_list), 1)
-            properties = self.client.capture.call_args_list[0][1]["properties"]
+            self.assertEqual(len(self.client._enqueue.call_args_list), 1)
+            properties = self.client._enqueue.call_args_list[0][0][0]["properties"]
             self._assert_system_prompt_captured(properties["$ai_input"])

hanzo_insights/test/ai/test_tokens_source.py

@@ -0,0 +1,62 @@
+from parameterized import parameterized
+
+from hanzo_insights.ai.utils import _get_tokens_source
+
+
+@parameterized.expand(
+    [
+        ("no_insights_properties", {"$ai_input_tokens": 100}, None, "sdk"),
+        ("empty_insights_properties", {"$ai_input_tokens": 100}, {}, "sdk"),
+        (
+            "unrelated_insights_properties",
+            {"$ai_input_tokens": 100},
+            {"foo": "bar"},
+            "sdk",
+        ),
+        (
+            "override_input_tokens",
+            {"$ai_input_tokens": 100},
+            {"$ai_input_tokens": 999},
+            "passthrough",
+        ),
+        (
+            "override_output_tokens",
+            {"$ai_output_tokens": 50},
+            {"$ai_output_tokens": 999},
+            "passthrough",
+        ),
+        (
+            "override_total_tokens",
+            {"$ai_input_tokens": 100},
+            {"$ai_total_tokens": 999},
+            "passthrough",
+        ),
+        (
+            "override_cache_read",
+            {"$ai_input_tokens": 100},
+            {"$ai_cache_read_input_tokens": 500},
+            "passthrough",
+        ),
+        (
+            "override_cache_creation",
+            {"$ai_input_tokens": 100},
+            {"$ai_cache_creation_input_tokens": 200},
+            "passthrough",
+        ),
+        (
+            "override_reasoning_tokens",
+            {"$ai_input_tokens": 100},
+            {"$ai_reasoning_tokens": 300},
+            "passthrough",
+        ),
+        (
+            "mixed_override_and_custom",
+            {"$ai_input_tokens": 100},
+            {"$ai_input_tokens": 999, "custom_key": "value"},
+            "passthrough",
+        ),
+    ]
+)
+def test_get_tokens_source(name, sdk_tags, insights_properties, expected):
+    result = _get_tokens_source(sdk_tags, insights_properties)
+    assert result == expected

hanzo_insights/test/integrations/test_middleware.py

@@ -1,4 +1,4 @@
-from posthog.contexts import (
+from hanzo_insights.contexts import (
     new_context,
     get_context_session_id,
     get_context_distinct_id,
@@ -20,7 +20,7 @@ if not settings.configured:
     )
     django.setup()
 
-from posthog.integrations.django import PosthogContextMiddleware
+from hanzo_insights.integrations.django import InsightsContextMiddleware
 
 
 class MockRequest:
@@ -45,7 +45,7 @@ class MockRequest:
         return f"{scheme}://{self._host}{self.path}"
 
 
-class TestPosthogContextMiddleware(unittest.TestCase):
+class TestInsightsContextMiddleware(unittest.TestCase):
     def create_middleware(
         self,
         extra_tags=None,
@@ -60,24 +60,24 @@ class TestPosthogContextMiddleware(unittest.TestCase):
 
         with patch("django.conf.settings") as mock_settings:
             # Configure mock settings
-            mock_settings.POSTHOG_MW_EXTRA_TAGS = extra_tags
-            mock_settings.POSTHOG_MW_REQUEST_FILTER = request_filter
-            mock_settings.POSTHOG_MW_TAG_MAP = tag_map
-            mock_settings.POSTHOG_MW_CAPTURE_EXCEPTIONS = capture_exceptions
-            mock_settings.POSTHOG_MW_CLIENT = None
+            mock_settings.INSIGHTS_MW_EXTRA_TAGS = extra_tags
+            mock_settings.INSIGHTS_MW_REQUEST_FILTER = request_filter
+            mock_settings.INSIGHTS_MW_TAG_MAP = tag_map
+            mock_settings.INSIGHTS_MW_CAPTURE_EXCEPTIONS = capture_exceptions
+            mock_settings.INSIGHTS_MW_CLIENT = None
 
             # Make hasattr work correctly
             def mock_hasattr(obj, name):
                 return name in [
-                    "POSTHOG_MW_EXTRA_TAGS",
-                    "POSTHOG_MW_REQUEST_FILTER",
-                    "POSTHOG_MW_TAG_MAP",
-                    "POSTHOG_MW_CAPTURE_EXCEPTIONS",
-                    "POSTHOG_MW_CLIENT",
+                    "INSIGHTS_MW_EXTRA_TAGS",
+                    "INSIGHTS_MW_REQUEST_FILTER",
+                    "INSIGHTS_MW_TAG_MAP",
+                    "INSIGHTS_MW_CAPTURE_EXCEPTIONS",
+                    "INSIGHTS_MW_CLIENT",
                 ]
 
             with patch("builtins.hasattr", side_effect=mock_hasattr):
-                middleware = PosthogContextMiddleware(get_response)
+                middleware = InsightsContextMiddleware(get_response)
 
         return middleware
 
@@ -87,8 +87,8 @@ class TestPosthogContextMiddleware(unittest.TestCase):
             middleware = self.create_middleware()
             request = MockRequest(
                 headers={
-                    "X-POSTHOG-SESSION-ID": "session-123",
-                    "X-POSTHOG-DISTINCT-ID": "user-456",
+                    "X-INSIGHTS-SESSION-ID": "session-123",
+                    "X-INSIGHTS-DISTINCT-ID": "user-456",
                 },
                 method="POST",
                 path="/api/test",
@@ -104,7 +104,7 @@ class TestPosthogContextMiddleware(unittest.TestCase):
             self.assertEqual(tags["$request_method"], "POST")
 
     def test_extract_tags_missing_headers(self):
-        """Test tag extraction when PostHog headers are missing"""
+        """Test tag extraction when Insights headers are missing"""
 
         with new_context():
             middleware = self.create_middleware()
@@ -118,12 +118,12 @@ class TestPosthogContextMiddleware(unittest.TestCase):
             self.assertEqual(tags["$request_method"], "GET")
 
     def test_extract_tags_partial_headers(self):
-        """Test tag extraction with only some PostHog headers present"""
+        """Test tag extraction with only some Insights headers present"""
 
         with new_context():
             middleware = self.create_middleware()
             request = MockRequest(
-                headers={"X-POSTHOG-SESSION-ID": "session-only"}, method="PUT"
+                headers={"X-INSIGHTS-SESSION-ID": "session-only"}, method="PUT"
             )
 
             tags = middleware.extract_tags(request)
@@ -141,7 +141,7 @@ class TestPosthogContextMiddleware(unittest.TestCase):
         with new_context():
             middleware = self.create_middleware(extra_tags=extra_tags_func)
             request = MockRequest(
-                headers={"X-POSTHOG-SESSION-ID": "session-123"}, method="GET"
+                headers={"X-INSIGHTS-SESSION-ID": "session-123"}, method="GET"
             )
 
             tags = middleware.extract_tags(request)
@@ -167,7 +167,7 @@ class TestPosthogContextMiddleware(unittest.TestCase):
                 tag_map=tag_map_func, extra_tags=extra_tags_func
             )
             request = MockRequest(
-                headers={"X-POSTHOG-SESSION-ID": "session-123"}, method="GET"
+                headers={"X-INSIGHTS-SESSION-ID": "session-123"}, method="GET"
             )
 
             tags = middleware.extract_tags(request)
@@ -230,7 +230,7 @@ class TestPosthogContextMiddleware(unittest.TestCase):
         middleware.client = mock_client
 
         request = MockRequest(
-            headers={"X-POSTHOG-DISTINCT-ID": "test-user"},
+            headers={"X-INSIGHTS-DISTINCT-ID": "test-user"},
             method="POST",
             path="/api/endpoint",
         )
@@ -282,7 +282,7 @@ class TestPosthogContextMiddleware(unittest.TestCase):
         mock_client.capture_exception.assert_not_called()
 
 
-class TestPosthogContextMiddlewareSync(unittest.TestCase):
+class TestInsightsContextMiddlewareSync(unittest.TestCase):
     """Test synchronous middleware behavior"""
 
     def test_sync_middleware_call(self):
@@ -291,13 +291,13 @@ class TestPosthogContextMiddlewareSync(unittest.TestCase):
         get_response = Mock(return_value=mock_response)
 
         # Create middleware with sync get_response
-        middleware = PosthogContextMiddleware(get_response)
+        middleware = InsightsContextMiddleware(get_response)
 
         # Verify sync mode detected
         self.assertFalse(middleware._is_coroutine)
 
         request = MockRequest(
-            headers={"X-POSTHOG-SESSION-ID": "test-session"},
+            headers={"X-INSIGHTS-SESSION-ID": "test-session"},
             method="GET",
             path="/test",
         )
@@ -318,7 +318,7 @@ class TestPosthogContextMiddlewareSync(unittest.TestCase):
         def request_filter(req):
             return False
 
-        middleware = PosthogContextMiddleware.__new__(PosthogContextMiddleware)
+        middleware = InsightsContextMiddleware.__new__(InsightsContextMiddleware)
         middleware.get_response = get_response
         middleware._is_coroutine = False
         middleware.request_filter = request_filter
@@ -351,7 +351,7 @@ class TestPosthogContextMiddlewareSync(unittest.TestCase):
         mock_client = Mock()
         get_response = Mock(return_value=Mock(status_code=500))
 
-        middleware = PosthogContextMiddleware(get_response)
+        middleware = InsightsContextMiddleware(get_response)
         middleware.client = mock_client
 
         def get_response_simulating_django(request):
@@ -380,7 +380,7 @@ class TestPosthogContextMiddlewareSync(unittest.TestCase):
             )
 
 
-class TestPosthogContextMiddlewareAsync(unittest.TestCase):
+class TestInsightsContextMiddlewareAsync(unittest.TestCase):
     """Test asynchronous middleware behavior"""
 
     def test_async_middleware_detection(self):
@@ -389,7 +389,7 @@ class TestPosthogContextMiddlewareAsync(unittest.TestCase):
         async def async_get_response(request):
             return Mock()
 
-        middleware = PosthogContextMiddleware(async_get_response)
+        middleware = InsightsContextMiddleware(async_get_response)
 
         # Verify async mode detected
         self.assertTrue(middleware._is_coroutine)
@@ -403,10 +403,10 @@ class TestPosthogContextMiddlewareAsync(unittest.TestCase):
             async def async_get_response(request):
                 return mock_response
 
-            middleware = PosthogContextMiddleware(async_get_response)
+            middleware = InsightsContextMiddleware(async_get_response)
 
             request = MockRequest(
-                headers={"X-POSTHOG-SESSION-ID": "async-session"},
+                headers={"X-INSIGHTS-SESSION-ID": "async-session"},
                 method="POST",
                 path="/async-test",
             )
@@ -434,7 +434,7 @@ class TestPosthogContextMiddlewareAsync(unittest.TestCase):
                 return mock_response
 
             # Properly initialize middleware
-            middleware = PosthogContextMiddleware(async_get_response)
+            middleware = InsightsContextMiddleware(async_get_response)
             # Override request filter after initialization
             middleware.request_filter = lambda req: False
 
@@ -459,10 +459,10 @@ class TestPosthogContextMiddlewareAsync(unittest.TestCase):
                 self.assertEqual(session_id, "async-session-123")
                 return mock_response
 
-            middleware = PosthogContextMiddleware(async_get_response)
+            middleware = InsightsContextMiddleware(async_get_response)
 
             request = MockRequest(
-                headers={"X-POSTHOG-SESSION-ID": "async-session-123"},
+                headers={"X-INSIGHTS-SESSION-ID": "async-session-123"},
                 method="GET",
             )
 
@@ -483,7 +483,7 @@ class TestPosthogContextMiddlewareAsync(unittest.TestCase):
                 raise ValueError("Async test exception")
 
             # Properly initialize middleware
-            middleware = PosthogContextMiddleware(raise_exception)
+            middleware = InsightsContextMiddleware(raise_exception)
             middleware.client = mock_client  # Override with mock client
 
             request = MockRequest()
@@ -525,11 +525,11 @@ class TestPosthogContextMiddlewareAsync(unittest.TestCase):
                 self.assertEqual(distinct_id, "123")
                 return mock_response
 
-            middleware = PosthogContextMiddleware(async_get_response)
+            middleware = InsightsContextMiddleware(async_get_response)
             middleware.client = Mock()
 
             request = MockRequest(
-                headers={"X-POSTHOG-SESSION-ID": "test-session"}, method="GET"
+                headers={"X-INSIGHTS-SESSION-ID": "test-session"}, method="GET"
             )
 
             # Mock auser() to return authenticated user
@@ -561,11 +561,11 @@ class TestPosthogContextMiddlewareAsync(unittest.TestCase):
                 self.assertIsNone(distinct_id)
                 return mock_response
 
-            middleware = PosthogContextMiddleware(async_get_response)
+            middleware = InsightsContextMiddleware(async_get_response)
             middleware.client = Mock()
 
             request = MockRequest(
-                headers={"X-POSTHOG-SESSION-ID": "test-session"}, method="GET"
+                headers={"X-INSIGHTS-SESSION-ID": "test-session"}, method="GET"
             )
 
             async def mock_auser():
@@ -591,12 +591,12 @@ class TestPosthogContextMiddlewareAsync(unittest.TestCase):
             async def async_get_response(request):
                 return mock_response
 
-            middleware = PosthogContextMiddleware(async_get_response)
+            middleware = InsightsContextMiddleware(async_get_response)
             middleware.client = Mock()
 
             # Request without auser method (no auth middleware)
             request = MockRequest(
-                headers={"X-POSTHOG-SESSION-ID": "test-session"}, method="GET"
+                headers={"X-INSIGHTS-SESSION-ID": "test-session"}, method="GET"
             )
 
             with new_context():
@@ -621,12 +621,12 @@ class TestPosthogContextMiddlewareAsync(unittest.TestCase):
             async def async_get_response(request):
                 return mock_response
 
-            middleware = PosthogContextMiddleware(async_get_response)
+            middleware = InsightsContextMiddleware(async_get_response)
             middleware.extra_tags = extra_tags_callback
             middleware.client = Mock()
 
             request = MockRequest(
-                headers={"X-POSTHOG-SESSION-ID": "test-session"}, method="GET"
+                headers={"X-INSIGHTS-SESSION-ID": "test-session"}, method="GET"
             )
 
             # Mock auser for no user
@@ -658,12 +658,12 @@ class TestPosthogContextMiddlewareAsync(unittest.TestCase):
             async def async_get_response(request):
                 return mock_response
 
-            middleware = PosthogContextMiddleware(async_get_response)
+            middleware = InsightsContextMiddleware(async_get_response)
             middleware.tag_map = tag_map_callback
             middleware.client = Mock()
 
             request = MockRequest(
-                headers={"X-POSTHOG-SESSION-ID": "test-session"}, method="GET"
+                headers={"X-INSIGHTS-SESSION-ID": "test-session"}, method="GET"
             )
 
             # Mock auser for no user
@@ -699,12 +699,12 @@ class TestPosthogContextMiddlewareAsync(unittest.TestCase):
                 self.assertEqual(session_id, "async-sess-123")
                 return mock_response
 
-            middleware = PosthogContextMiddleware(async_get_response)
+            middleware = InsightsContextMiddleware(async_get_response)
             middleware.client = Mock()
 
             request = MockRequest(
                 headers={
-                    "X-POSTHOG-SESSION-ID": "async-sess-123",
+                    "X-INSIGHTS-SESSION-ID": "async-sess-123",
                     "X-Forwarded-For": "192.168.1.1",
                     "User-Agent": "TestAgent/1.0",
                 },
@@ -725,13 +725,13 @@ class TestPosthogContextMiddlewareAsync(unittest.TestCase):
         asyncio.run(run_test())
 
 
-class TestPosthogContextMiddlewareHybrid(unittest.TestCase):
+class TestInsightsContextMiddlewareHybrid(unittest.TestCase):
     """Test hybrid middleware behavior with mixed sync/async chains"""
 
     def test_hybrid_flags_set(self):
         """Test that both capability flags are set"""
-        self.assertTrue(PosthogContextMiddleware.sync_capable)
-        self.assertTrue(PosthogContextMiddleware.async_capable)
+        self.assertTrue(InsightsContextMiddleware.sync_capable)
+        self.assertTrue(InsightsContextMiddleware.async_capable)
 
     def test_sync_to_async_routing(self):
         """Test that __call__ routes to __acall__ when async"""
@@ -740,7 +740,7 @@ class TestPosthogContextMiddlewareHybrid(unittest.TestCase):
             async def async_get_response(request):
                 return Mock()
 
-            middleware = PosthogContextMiddleware(async_get_response)
+            middleware = InsightsContextMiddleware(async_get_response)
 
             # Verify routing happens
             request = MockRequest()
@@ -759,7 +759,7 @@ class TestPosthogContextMiddlewareHybrid(unittest.TestCase):
         def sync_get_response(request):
             return mock_response
 
-        middleware = PosthogContextMiddleware(sync_get_response)
+        middleware = InsightsContextMiddleware(sync_get_response)
 
         request = MockRequest()
         result = middleware(request)

hanzo_insights/test/test_before_send.py

@@ -2,16 +2,16 @@ import unittest
 
 import mock
 
-from posthog.client import Client
-from posthog.test.test_utils import FAKE_TEST_API_KEY
+from hanzo_insights.client import Client
+from hanzo_insights.test.test_utils import FAKE_TEST_API_KEY
 
 
 class TestClient(unittest.TestCase):
     @classmethod
     def setUpClass(cls):
         # This ensures no real HTTP POST requests are made
-        cls.client_post_patcher = mock.patch("posthog.client.batch_post")
-        cls.consumer_post_patcher = mock.patch("posthog.consumer.batch_post")
+        cls.client_post_patcher = mock.patch("hanzo_insights.client.batch_post")
+        cls.consumer_post_patcher = mock.patch("hanzo_insights.consumer.batch_post")
         cls.client_post_patcher.start()
         cls.consumer_post_patcher.start()
 
@@ -40,7 +40,7 @@ class TestClient(unittest.TestCase):
             event["properties"]["processed_by_before_send"] = True
             return event
 
-        with mock.patch("posthog.client.batch_post") as mock_post:
+        with mock.patch("hanzo_insights.client.batch_post") as mock_post:
             client = Client(
                 FAKE_TEST_API_KEY,
                 on_error=self.set_fail,
@@ -73,7 +73,7 @@ class TestClient(unittest.TestCase):
                 return None
             return event
 
-        with mock.patch("posthog.client.batch_post") as mock_post:
+        with mock.patch("hanzo_insights.client.batch_post") as mock_post:
             client = Client(
                 FAKE_TEST_API_KEY,
                 on_error=self.set_fail,
@@ -101,7 +101,7 @@ class TestClient(unittest.TestCase):
         def buggy_before_send(event):
             raise ValueError("Oops!")
 
-        with mock.patch("posthog.client.batch_post") as mock_post:
+        with mock.patch("hanzo_insights.client.batch_post") as mock_post:
             client = Client(
                 FAKE_TEST_API_KEY,
                 on_error=self.set_fail,
@@ -128,7 +128,7 @@ class TestClient(unittest.TestCase):
             event["properties"]["marked"] = True
             return event
 
-        with mock.patch("posthog.client.batch_post") as mock_post:
+        with mock.patch("hanzo_insights.client.batch_post") as mock_post:
             client = Client(
                 FAKE_TEST_API_KEY,
                 on_error=self.set_fail,
@@ -153,7 +153,7 @@ class TestClient(unittest.TestCase):
 
     def test_before_send_callback_disabled_when_none(self):
         """Test that client works normally when before_send is None."""
-        with mock.patch("posthog.client.batch_post") as mock_post:
+        with mock.patch("hanzo_insights.client.batch_post") as mock_post:
             client = Client(
                 FAKE_TEST_API_KEY,
                 on_error=self.set_fail,
@@ -189,7 +189,7 @@ class TestClient(unittest.TestCase):
 
             return event
 
-        with mock.patch("posthog.client.batch_post") as mock_post:
+        with mock.patch("hanzo_insights.client.batch_post") as mock_post:
             client = Client(
                 FAKE_TEST_API_KEY,
                 on_error=self.set_fail,

hanzo_insights/test/test_consumer.py

@@ -0,0 +1,256 @@
+import json
+import time
+import unittest
+from typing import Any
+
+import mock
+from parameterized import parameterized
+
+try:
+    from queue import Queue
+except ImportError:
+    from Queue import Queue
+
+from hanzo_insights.consumer import MAX_MSG_SIZE, Consumer
+from hanzo_insights.request import APIError
+from hanzo_insights.test.test_utils import TEST_API_KEY
+
+
+def _track_event(event_name: str = "python event") -> dict[str, str]:
+    return {"type": "track", "event": event_name, "distinct_id": "distinct_id"}
+
+
+class TestConsumer(unittest.TestCase):
+    def test_next(self) -> None:
+        q = Queue()
+        consumer = Consumer(q, "")
+        q.put(1)
+        next = consumer.next()
+        self.assertEqual(next, [1])
+
+    def test_next_limit(self) -> None:
+        q = Queue()
+        flush_at = 50
+        consumer = Consumer(q, "", flush_at)
+        for i in range(10000):
+            q.put(i)
+        next = consumer.next()
+        self.assertEqual(next, list(range(flush_at)))
+
+    def test_dropping_oversize_msg(self) -> None:
+        q = Queue()
+        consumer = Consumer(q, "")
+        oversize_msg = {"m": "x" * MAX_MSG_SIZE}
+        q.put(oversize_msg)
+        next = consumer.next()
+        self.assertEqual(next, [])
+        self.assertTrue(q.empty())
+
+    def test_upload(self) -> None:
+        q = Queue()
+        consumer = Consumer(q, TEST_API_KEY)
+        q.put(_track_event())
+        success = consumer.upload()
+        self.assertTrue(success)
+
+    def test_flush_interval(self) -> None:
+        # Put _n_ items in the queue, pausing a little bit more than
+        # _flush_interval_ after each one.
+        # The consumer should upload _n_ times.
+        q = Queue()
+        flush_interval = 0.3
+        consumer = Consumer(q, TEST_API_KEY, flush_at=10, flush_interval=flush_interval)
+        with mock.patch("hanzo_insights.consumer.batch_post") as mock_post:
+            consumer.start()
+            for i in range(3):
+                q.put(_track_event("python event %d" % i))
+                time.sleep(flush_interval * 1.1)
+            self.assertEqual(mock_post.call_count, 3)
+
+    def test_multiple_uploads_per_interval(self) -> None:
+        # Put _flush_at*2_ items in the queue at once, then pause for
+        # _flush_interval_. The consumer should upload 2 times.
+        q = Queue()
+        flush_interval = 0.5
+        flush_at = 10
+        consumer = Consumer(
+            q, TEST_API_KEY, flush_at=flush_at, flush_interval=flush_interval
+        )
+        with mock.patch("hanzo_insights.consumer.batch_post") as mock_post:
+            consumer.start()
+            for i in range(flush_at * 2):
+                q.put(_track_event("python event %d" % i))
+            time.sleep(flush_interval * 1.1)
+            self.assertEqual(mock_post.call_count, 2)
+
+    def test_request(self) -> None:
+        consumer = Consumer(None, TEST_API_KEY)
+        consumer.request([_track_event()])
+
+    def _run_retry_test(
+        self, exception: Exception, exception_count: int, retries: int = 10
+    ) -> None:
+        call_count = [0]
+
+        def mock_post(*args: Any, **kwargs: Any) -> None:
+            call_count[0] += 1
+            if call_count[0] <= exception_count:
+                raise exception
+
+        consumer = Consumer(None, TEST_API_KEY, retries=retries)
+        with mock.patch(
+            "hanzo_insights.consumer.batch_post", mock.Mock(side_effect=mock_post)
+        ):
+            if exception_count <= retries:
+                consumer.request([_track_event()])
+            else:
+                with self.assertRaises(type(exception)):
+                    consumer.request([_track_event()])
+
+    @parameterized.expand(
+        [
+            ("general_errors", Exception("generic exception"), 2),
+            ("server_errors", APIError(500, "Internal Server Error"), 2),
+            ("rate_limit_errors", APIError(429, "Too Many Requests"), 2),
+        ]
+    )
+    def test_request_retries_on_retriable_errors(
+        self, _name: str, exception: Exception, exception_count: int
+    ) -> None:
+        self._run_retry_test(exception, exception_count)
+
+    def test_request_does_not_retry_client_errors(self) -> None:
+        with self.assertRaises(APIError):
+            self._run_retry_test(APIError(400, "Client Errors"), 1)
+
+    def test_request_fails_when_exceptions_exceed_retries(self) -> None:
+        self._run_retry_test(APIError(500, "Internal Server Error"), 4, retries=3)
+
+    def test_pause(self) -> None:
+        consumer = Consumer(None, TEST_API_KEY)
+        consumer.pause()
+        self.assertFalse(consumer.running)
+
+    def test_max_batch_size(self) -> None:
+        q = Queue()
+        consumer = Consumer(q, TEST_API_KEY, flush_at=100000, flush_interval=3)
+        properties = {}
+        for n in range(0, 500):
+            properties[str(n)] = "one_long_property_value_to_build_a_big_event"
+        track = {
+            "type": "track",
+            "event": "python event",
+            "distinct_id": "distinct_id",
+            "properties": properties,
+        }
+        msg_size = len(json.dumps(track).encode())
+        # Let's capture 8MB of data to trigger two batches
+        n_msgs = int(8_000_000 / msg_size)
+
+        def mock_post_fn(_: str, data: str, **kwargs: Any) -> mock.Mock:
+            res = mock.Mock()
+            res.status_code = 200
+            request_size = len(data.encode())
+            # Batches close after the first message bringing it bigger than BATCH_SIZE_LIMIT, let's add 10% of margin
+            self.assertTrue(
+                request_size < (5 * 1024 * 1024) * 1.1,
+                "batch size (%d) higher than limit" % request_size,
+            )
+            return res
+
+        with mock.patch(
+            "hanzo_insights.request._session.post", side_effect=mock_post_fn
+        ) as mock_post:
+            consumer.start()
+            for _ in range(0, n_msgs + 2):
+                q.put(track)
+            q.join()
+            self.assertEqual(mock_post.call_count, 2)
+
+    def test_request_sleeps_with_retry_after(self) -> None:
+        error = APIError(429, "Too Many Requests", retry_after=5.0)
+        call_count = [0]
+
+        def mock_post(*args: Any, **kwargs: Any) -> None:
+            call_count[0] += 1
+            if call_count[0] <= 1:
+                raise error
+
+        consumer = Consumer(None, TEST_API_KEY, retries=3)
+        with (
+            mock.patch("hanzo_insights.consumer.batch_post", side_effect=mock_post),
+            mock.patch("hanzo_insights.consumer.time.sleep") as mock_sleep,
+        ):
+            consumer.request([_track_event()])
+            mock_sleep.assert_called_once_with(5.0)
+
+    def test_request_uses_exponential_backoff_without_retry_after(self) -> None:
+        error = APIError(503, "Service Unavailable")
+        call_count = [0]
+
+        def mock_post(*args: Any, **kwargs: Any) -> None:
+            call_count[0] += 1
+            if call_count[0] <= 3:
+                raise error
+
+        consumer = Consumer(None, TEST_API_KEY, retries=3)
+        with (
+            mock.patch("hanzo_insights.consumer.batch_post", side_effect=mock_post),
+            mock.patch("hanzo_insights.consumer.time.sleep") as mock_sleep,
+        ):
+            consumer.request([_track_event()])
+            self.assertEqual(
+                mock_sleep.call_args_list,
+                [
+                    mock.call(1),  # 2^0
+                    mock.call(2),  # 2^1
+                    mock.call(4),  # 2^2
+                ],
+            )
+
+    def test_request_retries_on_408(self) -> None:
+        call_count = [0]
+
+        def mock_post(*args: Any, **kwargs: Any) -> None:
+            call_count[0] += 1
+            if call_count[0] <= 1:
+                raise APIError(408, "Request Timeout")
+
+        consumer = Consumer(None, TEST_API_KEY, retries=3)
+        with (
+            mock.patch("hanzo_insights.consumer.batch_post", side_effect=mock_post),
+            mock.patch("hanzo_insights.consumer.time.sleep"),
+        ):
+            consumer.request([_track_event()])
+            self.assertEqual(call_count[0], 2)
+
+    @parameterized.expand(
+        [
+            ("on_error_succeeds", False),
+            ("on_error_raises", True),
+        ]
+    )
+    def test_upload_exception_calls_on_error_and_does_not_raise(
+        self, _name: str, on_error_raises: bool
+    ) -> None:
+        on_error_called: list[tuple[Exception, list[dict[str, str]]]] = []
+
+        def on_error(e: Exception, batch: list[dict[str, str]]) -> None:
+            on_error_called.append((e, batch))
+            if on_error_raises:
+                raise Exception("on_error failed")
+
+        q = Queue()
+        consumer = Consumer(q, TEST_API_KEY, on_error=on_error)
+        track = _track_event()
+        q.put(track)
+
+        with mock.patch.object(
+            consumer, "request", side_effect=Exception("request failed")
+        ):
+            result = consumer.upload()
+
+        self.assertFalse(result)
+        self.assertEqual(len(on_error_called), 1)
+        self.assertEqual(str(on_error_called[0][0]), "request failed")
+        self.assertEqual(on_error_called[0][1], [track])

hanzo_insights/test/test_contexts.py

@@ -1,7 +1,7 @@
 import unittest
 from unittest.mock import patch
 
-from posthog.contexts import (
+from hanzo_insights.contexts import (
     get_tags,
     new_context,
     scoped,
@@ -66,7 +66,7 @@ class TestContexts(unittest.TestCase):
             # Back to level 1
             assert get_tags() == {"level1": "value1"}
 
-    @patch("posthog.capture_exception")
+    @patch("hanzo_insights.capture_exception")
     def test_scoped_decorator_success(self, mock_capture):
         @scoped()
         def successful_function(x, y):
@@ -85,7 +85,7 @@ class TestContexts(unittest.TestCase):
         # Context should be cleared after function execution
         assert get_tags() == {}
 
-    @patch("posthog.capture_exception")
+    @patch("hanzo_insights.capture_exception")
     def test_scoped_decorator_exception(self, mock_capture):
         test_exception = ValueError("Test exception")
 
@@ -111,7 +111,7 @@ class TestContexts(unittest.TestCase):
         # Context should be cleared after function execution
         assert get_tags() == {}
 
-    @patch("posthog.capture_exception")
+    @patch("hanzo_insights.capture_exception")
     def test_new_context_exception_handling(self, mock_capture):
         test_exception = RuntimeError("Context exception")
 
@@ -191,6 +191,32 @@ class TestContexts(unittest.TestCase):
             assert get_context_distinct_id() == "user123"
             assert get_context_session_id() == "session456"
 
+    def test_child_tags_override_parent_tags_in_non_fresh_context(self):
+        with new_context(fresh=True):
+            tag("shared_key", "parent_value")
+            tag("parent_only", "parent")
+
+            with new_context(fresh=False):
+                # Child should inherit parent tags
+                assert get_tags()["parent_only"] == "parent"
+
+                # Child sets same key - should override parent
+                tag("shared_key", "child_value")
+                tag("child_only", "child")
+
+                tags = get_tags()
+                # Child value should win for shared key
+                assert tags["shared_key"] == "child_value"
+                # Both parent and child tags should be present
+                assert tags["parent_only"] == "parent"
+                assert tags["child_only"] == "child"
+
+            # Parent context should be unchanged
+            parent_tags = get_tags()
+            assert parent_tags["shared_key"] == "parent_value"
+            assert parent_tags["parent_only"] == "parent"
+            assert "child_only" not in parent_tags
+
     def test_scoped_decorator_with_context_ids(self):
         @scoped()
         def function_with_context():

hanzo_insights/test/test_exception_capture.py

@@ -10,8 +10,8 @@ def test_excepthook(tmpdir):
     app.write(
         dedent(
             """
-    from posthog import Posthog
-    posthog = Posthog('phc_x', host='https://eu.i.posthog.com', enable_exception_autocapture=True, debug=True, on_error=lambda e, batch: print('error handling batch: ', e, batch))
+    from hanzo_insights import Insights
+    client = Insights('phc_x', host='https://eu.i.insights.hanzo.ai', enable_exception_autocapture=True, debug=True, on_error=lambda e, batch: print('error handling batch: ', e, batch))
 
     # frame_value = "LOL"
 
@@ -27,7 +27,7 @@ def test_excepthook(tmpdir):
 
     assert b"ZeroDivisionError" in output
     assert b"LOL" in output
-    assert b"DEBUG:posthog:data uploaded successfully" in output
+    assert b"DEBUG:hanzo_insights:data uploaded successfully" in output
     assert (
         b'"$exception_list": [{"mechanism": {"type": "generic", "handled": true}, "module": null, "type": "ZeroDivisionError", "value": "division by zero", "stacktrace": {"frames": [{"platform": "python", "filename": "app.py", "abs_path"'
         in output
@@ -40,14 +40,14 @@ def test_code_variables_capture(tmpdir):
         dedent(
             """
     import os
-    from posthog import Posthog
+    from hanzo_insights import Insights
     
     class UnserializableObject:
         pass
     
-    posthog = Posthog(
+    client = Insights(
         'phc_x', 
-        host='https://eu.i.posthog.com', 
+        host='https://eu.i.insights.hanzo.ai', 
         debug=True, 
         enable_exception_autocapture=True,
         capture_exception_code_variables=True,
@@ -118,29 +118,29 @@ def test_code_variables_capture(tmpdir):
     assert b"'my_bool': 'True'" in output
     assert b'"my_dict": "{\\"name\\": \\"test\\", \\"value\\": 123}"' in output
     assert (
-        b'{\\"safe_key\\": \\"safe_value\\", \\"password\\": \\"$$_posthog_redacted_based_on_masking_rules_$$\\", \\"other_key\\": \\"$$_posthog_redacted_based_on_masking_rules_$$\\"}'
+        b'{\\"safe_key\\": \\"safe_value\\", \\"password\\": \\"$$_insights_redacted_based_on_masking_rules_$$\\", \\"other_key\\": \\"$$_insights_redacted_based_on_masking_rules_$$\\"}'
         in output
     )
     assert (
-        b'{\\"level1\\": {\\"level2\\": {\\"api_key\\": \\"$$_posthog_redacted_based_on_masking_rules_$$\\", \\"data\\": \\"$$_posthog_redacted_based_on_masking_rules_$$\\", \\"safe\\": \\"visible\\"}}}'
+        b'{\\"level1\\": {\\"level2\\": {\\"api_key\\": \\"$$_insights_redacted_based_on_masking_rules_$$\\", \\"data\\": \\"$$_insights_redacted_based_on_masking_rules_$$\\", \\"safe\\": \\"visible\\"}}}'
         in output
     )
     assert (
-        b'[\\"safe_item\\", \\"$$_posthog_redacted_based_on_masking_rules_$$\\", \\"another_safe\\"]'
+        b'[\\"safe_item\\", \\"$$_insights_redacted_based_on_masking_rules_$$\\", \\"another_safe\\"]'
         in output
     )
     assert (
-        b'[\\"tuple_safe\\", \\"$$_posthog_redacted_based_on_masking_rules_$$\\", \\"tuple_also_safe\\"]'
+        b'[\\"tuple_safe\\", \\"$$_insights_redacted_based_on_masking_rules_$$\\", \\"tuple_also_safe\\"]'
         in output
     )
     assert (
-        b'[{\\"id\\": 1, \\"password\\": \\"$$_posthog_redacted_based_on_masking_rules_$$\\"}, {\\"id\\": 2, \\"value\\": \\"safe_value\\"}]'
+        b'[{\\"id\\": 1, \\"password\\": \\"$$_insights_redacted_based_on_masking_rules_$$\\"}, {\\"id\\": 2, \\"value\\": \\"safe_value\\"}]'
         in output
     )
     assert b"<__main__.UnserializableObject object at" in output
-    assert b"'my_password': '$$_posthog_redacted_based_on_masking_rules_$$'" in output
+    assert b"'my_password': '$$_insights_redacted_based_on_masking_rules_$$'" in output
     assert (
-        b"'my_innocent_var': '$$_posthog_redacted_based_on_masking_rules_$$'" in output
+        b"'my_innocent_var': '$$_insights_redacted_based_on_masking_rules_$$'" in output
     )
     assert b"'__should_be_ignored':" not in output
 
@@ -160,12 +160,12 @@ def test_code_variables_context_override(tmpdir):
         dedent(
             """
     import os
-    import posthog
-    from posthog import Posthog
+    import hanzo_insights
+    from hanzo_insights import Insights
     
-    posthog_client = Posthog(
+    insights_client = Insights(
         'phc_x', 
-        host='https://eu.i.posthog.com', 
+        host='https://eu.i.insights.hanzo.ai', 
         debug=True, 
         enable_exception_autocapture=True,
         capture_exception_code_variables=False,
@@ -178,10 +178,10 @@ def test_code_variables_context_override(tmpdir):
         
         1/0
     
-    with posthog.new_context(client=posthog_client):
-        posthog.set_capture_exception_code_variables_context(True)
-        posthog.set_code_variables_mask_patterns_context([r"(?i).*bank.*"])
-        posthog.set_code_variables_ignore_patterns_context([])
+    with hanzo_insights.new_context(client=insights_client):
+        hanzo_insights.set_capture_exception_code_variables_context(True)
+        hanzo_insights.set_code_variables_mask_patterns_context([r"(?i).*bank.*"])
+        hanzo_insights.set_code_variables_ignore_patterns_context([])
         
         process_data()
     """
@@ -195,7 +195,7 @@ def test_code_variables_context_override(tmpdir):
 
     assert b"ZeroDivisionError" in output
     assert b"code_variables" in output
-    assert b"'bank': '$$_posthog_redacted_based_on_masking_rules_$$'" in output
+    assert b"'bank': '$$_insights_redacted_based_on_masking_rules_$$'" in output
     assert b"'__dunder_var': 'should_be_visible'" in output
 
 
@@ -205,11 +205,11 @@ def test_code_variables_size_limiter(tmpdir):
         dedent(
             """
     import os
-    from posthog import Posthog
+    from hanzo_insights import Insights
     
-    posthog = Posthog(
+    client = Insights(
         'phc_x', 
-        host='https://eu.i.posthog.com', 
+        host='https://eu.i.insights.hanzo.ai', 
         debug=True, 
         enable_exception_autocapture=True,
         capture_exception_code_variables=True,
@@ -299,11 +299,11 @@ def test_code_variables_disabled_capture(tmpdir):
         dedent(
             """
     import os
-    from posthog import Posthog
+    from hanzo_insights import Insights
     
-    posthog = Posthog(
+    client = Insights(
         'phc_x', 
-        host='https://eu.i.posthog.com', 
+        host='https://eu.i.insights.hanzo.ai', 
         debug=True, 
         enable_exception_autocapture=True,
         capture_exception_code_variables=False,
@@ -340,12 +340,12 @@ def test_code_variables_enabled_then_disabled_in_context(tmpdir):
         dedent(
             """
     import os
-    import posthog
-    from posthog import Posthog
+    import hanzo_insights
+    from hanzo_insights import Insights
     
-    posthog_client = Posthog(
+    insights_client = Insights(
         'phc_x', 
-        host='https://eu.i.posthog.com', 
+        host='https://eu.i.insights.hanzo.ai', 
         debug=True, 
         enable_exception_autocapture=True,
         capture_exception_code_variables=True,
@@ -358,8 +358,8 @@ def test_code_variables_enabled_then_disabled_in_context(tmpdir):
         
         1/0
     
-    with posthog.new_context(client=posthog_client):
-        posthog.set_capture_exception_code_variables_context(False)
+    with hanzo_insights.new_context(client=insights_client):
+        hanzo_insights.set_capture_exception_code_variables_context(False)
         
         process_data()
     """
@@ -388,15 +388,15 @@ def test_code_variables_repr_fallback(tmpdir):
     from datetime import datetime, timedelta
     from decimal import Decimal
     from fractions import Fraction
-    from posthog import Posthog
+    from hanzo_insights import Insights
     
     class CustomReprClass:
         def __repr__(self):
             return '<CustomReprClass: custom representation>'
     
-    posthog = Posthog(
+    client = Insights(
         'phc_x', 
-        host='https://eu.i.posthog.com', 
+        host='https://eu.i.insights.hanzo.ai', 
         debug=True, 
         enable_exception_autocapture=True,
         capture_exception_code_variables=True,
@@ -450,3 +450,240 @@ def test_code_variables_repr_fallback(tmpdir):
     assert "<CustomReprClass: custom representation>" in output
     assert "<lambda>" in output
     assert "<function trigger_error at" in output
+
+
+def test_code_variables_too_long_string_value_replaced(tmpdir):
+    app = tmpdir.join("app.py")
+    app.write(
+        dedent(
+            """
+    import os
+    from hanzo_insights import Insights
+
+    client = Insights(
+        'phc_x',
+        host='https://eu.i.insights.hanzo.ai',
+        debug=True,
+        enable_exception_autocapture=True,
+        capture_exception_code_variables=True,
+        project_root=os.path.dirname(os.path.abspath(__file__))
+    )
+
+    def trigger_error():
+        short_value = "I am short"
+        long_value = "x" * 20000
+        long_blob = "password_" + "a" * 20000
+
+        1/0
+
+    trigger_error()
+    """
+        )
+    )
+
+    with pytest.raises(subprocess.CalledProcessError) as excinfo:
+        subprocess.check_output([sys.executable, str(app)], stderr=subprocess.STDOUT)
+
+    output = excinfo.value.output.decode("utf-8")
+
+    assert "ZeroDivisionError" in output
+    assert "code_variables" in output
+
+    assert "'short_value': 'I am short'" in output
+
+    assert "$$_insights_value_too_long_$$" in output
+
+    assert "'long_blob': '$$_insights_value_too_long_$$'" in output
+
+
+def test_code_variables_too_long_string_in_nested_dict(tmpdir):
+    app = tmpdir.join("app.py")
+    app.write(
+        dedent(
+            """
+    import os
+    from hanzo_insights import Insights
+
+    client = Insights(
+        'phc_x',
+        host='https://eu.i.insights.hanzo.ai',
+        debug=True,
+        enable_exception_autocapture=True,
+        capture_exception_code_variables=True,
+        project_root=os.path.dirname(os.path.abspath(__file__))
+    )
+
+    def trigger_error():
+        my_data = {
+            "short_key": "short_val",
+            "long_key": "y" * 20000,
+            "nested": {
+                "deep_long": "z" * 20000,
+                "deep_short": "ok",
+            },
+        }
+
+        1/0
+
+    trigger_error()
+    """
+        )
+    )
+
+    with pytest.raises(subprocess.CalledProcessError) as excinfo:
+        subprocess.check_output([sys.executable, str(app)], stderr=subprocess.STDOUT)
+
+    output = excinfo.value.output.decode("utf-8")
+
+    assert "ZeroDivisionError" in output
+    assert "code_variables" in output
+
+    assert "short_val" in output
+    assert "ok" in output
+
+    assert "$$_insights_value_too_long_$$" in output
+    assert "y" * 1000 not in output
+    assert "z" * 1000 not in output
+
+
+def test_mask_sensitive_data_too_long_dict_key():
+    from hanzo_insights.exception_utils import (
+        CODE_VARIABLES_TOO_LONG_VALUE,
+        _compile_patterns,
+        _mask_sensitive_data,
+    )
+
+    compiled_mask = _compile_patterns([r"(?i)password"])
+
+    result = _mask_sensitive_data(
+        {
+            "short": "visible",
+            "k" * 20000: "hidden_val",
+            "password": "secret",
+        },
+        compiled_mask,
+    )
+
+    assert result["short"] == "visible"
+    # This then gets shortened by the JSON truncation at 1024 chars anyways so no worries
+    assert result["k" * 20000] == CODE_VARIABLES_TOO_LONG_VALUE
+    assert result["password"] == "$$_insights_redacted_based_on_masking_rules_$$"
+
+
+def test_mask_sensitive_data_circular_ref():
+    from hanzo_insights.exception_utils import _compile_patterns, _mask_sensitive_data
+
+    compiled_mask = _compile_patterns([r"(?i)password"])
+
+    # Circular dict
+    circular_dict = {"key": "value"}
+    circular_dict["self"] = circular_dict
+
+    result = _mask_sensitive_data(circular_dict, compiled_mask)
+    assert result["key"] == "value"
+    assert result["self"] == "<circular ref>"
+
+    # Circular list
+    circular_list = ["item"]
+    circular_list.append(circular_list)
+
+    result = _mask_sensitive_data(circular_list, compiled_mask)
+    assert result[0] == "item"
+    assert result[1] == "<circular ref>"
+
+
+def test_compile_patterns_fast_path_and_regex_fallback():
+    from hanzo_insights.exception_utils import _compile_patterns, _pattern_matches
+
+    # Simple case-insensitive patterns should become substrings
+    simple_only = _compile_patterns([r"(?i)password", r"(?i)token", r"(?i)jwt"])
+    substrings, regexes = simple_only
+    assert substrings == ["password", "token", "jwt"]
+    assert regexes == []
+
+    assert _pattern_matches("my_password_var", simple_only) is True
+    assert _pattern_matches("MY_TOKEN", simple_only) is True
+    assert _pattern_matches("safe_variable", simple_only) is False
+
+    # Complex regex patterns should stay as compiled regexes
+    complex_only = _compile_patterns([r"^__.*", r"\d{3,}", r"^sk_live_"])
+    substrings, regexes = complex_only
+    assert substrings == []
+    assert len(regexes) == 3
+
+    assert _pattern_matches("__dunder", complex_only) is True
+    assert _pattern_matches("has_999_numbers", complex_only) is True
+    assert _pattern_matches("sk_live_abc123", complex_only) is True
+    assert _pattern_matches("normal_var", complex_only) is False
+
+    # Mixed: simple substrings + complex regexes together
+    mixed = _compile_patterns(
+        [
+            r"(?i)secret",  # simple
+            r"(?i)api_key",  # simple
+            r"^__.*",  # regex
+            r"\btoken_\w+",  # regex
+        ]
+    )
+    substrings, regexes = mixed
+    assert substrings == ["secret", "api_key"]
+    assert len(regexes) == 2
+
+    # Substring matches
+    assert _pattern_matches("my_secret", mixed) is True
+    assert _pattern_matches("API_KEY_VALUE", mixed) is True
+
+    # Regex matches
+    assert _pattern_matches("__private", mixed) is True
+    assert _pattern_matches("token_abc", mixed) is True
+
+    # No match
+    assert _pattern_matches("safe_var", mixed) is False
+
+
+def test_mask_sensitive_data_large_dict_replaced():
+    from hanzo_insights.exception_utils import (
+        CODE_VARIABLES_TOO_LONG_VALUE,
+        _compile_patterns,
+        _mask_sensitive_data,
+    )
+
+    compiled_mask = _compile_patterns([r"(?i)password"])
+
+    large_dict = {f"key_{i}": f"value_{i}" for i in range(300)}
+
+    result = _mask_sensitive_data(large_dict, compiled_mask)
+
+    assert result == CODE_VARIABLES_TOO_LONG_VALUE
+
+
+def test_mask_sensitive_data_large_list_replaced():
+    from hanzo_insights.exception_utils import (
+        CODE_VARIABLES_TOO_LONG_VALUE,
+        _compile_patterns,
+        _mask_sensitive_data,
+    )
+
+    compiled_mask = _compile_patterns([r"(?i)password"])
+
+    large_list = [f"item_{i}" for i in range(300)]
+
+    result = _mask_sensitive_data(large_list, compiled_mask)
+
+    assert result == CODE_VARIABLES_TOO_LONG_VALUE
+
+
+def test_mask_sensitive_data_large_tuple_replaced():
+    from hanzo_insights.exception_utils import (
+        CODE_VARIABLES_TOO_LONG_VALUE,
+        _compile_patterns,
+        _mask_sensitive_data,
+    )
+
+    compiled_mask = _compile_patterns([r"(?i)password"])
+
+    large_tuple = tuple(f"item_{i}" for i in range(300))
+
+    result = _mask_sensitive_data(large_tuple, compiled_mask)
+
+    assert result == CODE_VARIABLES_TOO_LONG_VALUE

hanzo_insights/test/test_feature_flag.py

@@ -1,6 +1,6 @@
 import unittest
 
-from posthog.types import FeatureFlag, FlagMetadata, FlagReason, LegacyFlagMetadata
+from hanzo_insights.types import FeatureFlag, FlagMetadata, FlagReason, LegacyFlagMetadata
 
 
 class TestFeatureFlag(unittest.TestCase):

hanzo_insights/test/test_feature_flag_result.py

@@ -2,9 +2,9 @@ import unittest
 
 import mock
 
-from posthog.client import Client
-from posthog.test.test_utils import FAKE_TEST_API_KEY
-from posthog.types import (
+from hanzo_insights.client import Client
+from hanzo_insights.test.test_utils import FAKE_TEST_API_KEY
+from hanzo_insights.types import (
     FeatureFlag,
     FeatureFlagError,
     FeatureFlagResult,
@@ -328,7 +328,7 @@ class TestGetFeatureFlagResult(unittest.TestCase):
             disable_geoip=None,
         )
 
-    @mock.patch("posthog.client.flags")
+    @mock.patch("hanzo_insights.client.flags")
     @mock.patch.object(Client, "capture")
     def test_get_feature_flag_result_boolean_decide(self, patch_capture, patch_flags):
         patch_flags.return_value = {
@@ -375,7 +375,7 @@ class TestGetFeatureFlagResult(unittest.TestCase):
         captured_properties = patch_capture.call_args[1]["properties"]
         self.assertNotIn("$feature_flag_error", captured_properties)
 
-    @mock.patch("posthog.client.flags")
+    @mock.patch("hanzo_insights.client.flags")
     @mock.patch.object(Client, "capture")
     def test_get_feature_flag_result_variant_decide(self, patch_capture, patch_flags):
         patch_flags.return_value = {
@@ -421,7 +421,7 @@ class TestGetFeatureFlagResult(unittest.TestCase):
         captured_properties = patch_capture.call_args[1]["properties"]
         self.assertNotIn("$feature_flag_error", captured_properties)
 
-    @mock.patch("posthog.client.flags")
+    @mock.patch("hanzo_insights.client.flags")
     @mock.patch.object(Client, "capture")
     def test_get_feature_flag_result_unknown_flag(self, patch_capture, patch_flags):
         patch_flags.return_value = {
@@ -461,7 +461,7 @@ class TestGetFeatureFlagResult(unittest.TestCase):
             disable_geoip=None,
         )
 
-    @mock.patch("posthog.client.flags")
+    @mock.patch("hanzo_insights.client.flags")
     @mock.patch.object(Client, "capture")
     def test_get_feature_flag_result_with_errors_while_computing_flags(
         self, patch_capture, patch_flags
@@ -507,7 +507,7 @@ class TestGetFeatureFlagResult(unittest.TestCase):
             disable_geoip=None,
         )
 
-    @mock.patch("posthog.client.flags")
+    @mock.patch("hanzo_insights.client.flags")
     @mock.patch.object(Client, "capture")
     def test_get_feature_flag_result_flag_not_in_response(
         self, patch_capture, patch_flags
@@ -549,7 +549,7 @@ class TestGetFeatureFlagResult(unittest.TestCase):
             disable_geoip=None,
         )
 
-    @mock.patch("posthog.client.flags")
+    @mock.patch("hanzo_insights.client.flags")
     @mock.patch.object(Client, "capture")
     def test_get_feature_flag_result_errors_computing_and_flag_missing(
         self, patch_capture, patch_flags
@@ -585,7 +585,7 @@ class TestGetFeatureFlagResult(unittest.TestCase):
             disable_geoip=None,
         )
 
-    @mock.patch("posthog.client.flags")
+    @mock.patch("hanzo_insights.client.flags")
     @mock.patch.object(Client, "capture")
     def test_get_feature_flag_result_unknown_error(self, patch_capture, patch_flags):
         """Test that unexpected exceptions are captured as unknown_error."""
@@ -608,11 +608,11 @@ class TestGetFeatureFlagResult(unittest.TestCase):
             disable_geoip=None,
         )
 
-    @mock.patch("posthog.client.flags")
+    @mock.patch("hanzo_insights.client.flags")
     @mock.patch.object(Client, "capture")
     def test_get_feature_flag_result_timeout_error(self, patch_capture, patch_flags):
         """Test that timeout errors are captured specifically."""
-        from posthog.request import RequestsTimeout
+        from hanzo_insights.request import RequestsTimeout
 
         patch_flags.side_effect = RequestsTimeout("Request timed out")
 
@@ -633,11 +633,11 @@ class TestGetFeatureFlagResult(unittest.TestCase):
             disable_geoip=None,
         )
 
-    @mock.patch("posthog.client.flags")
+    @mock.patch("hanzo_insights.client.flags")
     @mock.patch.object(Client, "capture")
     def test_get_feature_flag_result_connection_error(self, patch_capture, patch_flags):
         """Test that connection errors are captured specifically."""
-        from posthog.request import RequestsConnectionError
+        from hanzo_insights.request import RequestsConnectionError
 
         patch_flags.side_effect = RequestsConnectionError("Connection refused")
 
@@ -658,11 +658,11 @@ class TestGetFeatureFlagResult(unittest.TestCase):
             disable_geoip=None,
         )
 
-    @mock.patch("posthog.client.flags")
+    @mock.patch("hanzo_insights.client.flags")
     @mock.patch.object(Client, "capture")
     def test_get_feature_flag_result_api_error(self, patch_capture, patch_flags):
         """Test that API errors include the status code."""
-        from posthog.request import APIError
+        from hanzo_insights.request import APIError
 
         patch_flags.side_effect = APIError(500, "Internal server error")
 
@@ -683,11 +683,11 @@ class TestGetFeatureFlagResult(unittest.TestCase):
             disable_geoip=None,
         )
 
-    @mock.patch("posthog.client.flags")
+    @mock.patch("hanzo_insights.client.flags")
     @mock.patch.object(Client, "capture")
     def test_get_feature_flag_result_quota_limited(self, patch_capture, patch_flags):
         """Test that quota limit errors are captured specifically."""
-        from posthog.request import QuotaLimitError
+        from hanzo_insights.request import QuotaLimitError
 
         patch_flags.side_effect = QuotaLimitError(429, "Rate limit exceeded")
 
@@ -712,7 +712,7 @@ class TestGetFeatureFlagResult(unittest.TestCase):
 class TestFeatureFlagErrorWithStaleCacheFallback(unittest.TestCase):
     """Tests for stale cache fallback behavior when flag evaluation fails.
 
-    When the PostHog API is unavailable (timeout, connection error, etc.), the SDK
+    When the Insights API is unavailable (timeout, connection error, etc.), the SDK
     falls back to stale cached flag values if available. These tests verify that:
     1. The stale cached value is returned when an error occurs
     2. The $feature_flag_error property is still set (for debugging)
@@ -741,11 +741,11 @@ class TestFeatureFlagErrorWithStaleCacheFallback(unittest.TestCase):
             flag_definition_version=self.client.flag_definition_version,
         )
 
-    @mock.patch("posthog.client.flags")
+    @mock.patch("hanzo_insights.client.flags")
     @mock.patch.object(Client, "capture")
     def test_timeout_error_returns_stale_cached_value(self, patch_capture, patch_flags):
         """Test that timeout errors return stale cached value when available."""
-        from posthog.request import RequestsTimeout
+        from hanzo_insights.request import RequestsTimeout
 
         # Pre-populate cache with a flag result
         cached_result = FeatureFlagResult.from_value_and_payload(
@@ -779,13 +779,13 @@ class TestFeatureFlagErrorWithStaleCacheFallback(unittest.TestCase):
             disable_geoip=None,
         )
 
-    @mock.patch("posthog.client.flags")
+    @mock.patch("hanzo_insights.client.flags")
     @mock.patch.object(Client, "capture")
     def test_connection_error_returns_stale_cached_value(
         self, patch_capture, patch_flags
     ):
         """Test that connection errors return stale cached value when available."""
-        from posthog.request import RequestsConnectionError
+        from hanzo_insights.request import RequestsConnectionError
 
         # Pre-populate cache with a boolean flag result
         cached_result = FeatureFlagResult.from_value_and_payload("my-flag", True, None)
@@ -816,11 +816,11 @@ class TestFeatureFlagErrorWithStaleCacheFallback(unittest.TestCase):
             disable_geoip=None,
         )
 
-    @mock.patch("posthog.client.flags")
+    @mock.patch("hanzo_insights.client.flags")
     @mock.patch.object(Client, "capture")
     def test_api_error_returns_stale_cached_value(self, patch_capture, patch_flags):
         """Test that API errors return stale cached value when available."""
-        from posthog.request import APIError
+        from hanzo_insights.request import APIError
 
         # Pre-populate cache
         cached_result = FeatureFlagResult.from_value_and_payload(
@@ -852,11 +852,11 @@ class TestFeatureFlagErrorWithStaleCacheFallback(unittest.TestCase):
             disable_geoip=None,
         )
 
-    @mock.patch("posthog.client.flags")
+    @mock.patch("hanzo_insights.client.flags")
     @mock.patch.object(Client, "capture")
     def test_error_without_cache_returns_none(self, patch_capture, patch_flags):
         """Test that errors return None when no stale cache is available."""
-        from posthog.request import RequestsTimeout
+        from hanzo_insights.request import RequestsTimeout
 
         # Do NOT populate cache - no fallback available
 

hanzo_insights/test/test_flag_definition_cache.py

@@ -1,7 +1,7 @@
 """
 Tests for FlagDefinitionCacheProvider functionality.
 
-These tests follow the patterns from the TypeScript implementation in posthog-js/packages/node.
+These tests follow the patterns from the TypeScript implementation in insights-js/packages/node.
 """
 
 import threading
@@ -9,13 +9,13 @@ import unittest
 from typing import Optional
 from unittest import mock
 
-from posthog.client import Client
-from posthog.flag_definition_cache import (
+from hanzo_insights.client import Client
+from hanzo_insights.flag_definition_cache import (
     FlagDefinitionCacheData,
     FlagDefinitionCacheProvider,
 )
-from posthog.request import GetResponse
-from posthog.test.test_utils import FAKE_TEST_API_KEY
+from hanzo_insights.request import GetResponse
+from hanzo_insights.test.test_utils import FAKE_TEST_API_KEY
 
 
 class MockCacheProvider:
@@ -63,8 +63,8 @@ class TestFlagDefinitionCacheProvider(unittest.TestCase):
     @classmethod
     def setUpClass(cls):
         # Prevent real HTTP requests
-        cls.client_post_patcher = mock.patch("posthog.client.batch_post")
-        cls.consumer_post_patcher = mock.patch("posthog.consumer.batch_post")
+        cls.client_post_patcher = mock.patch("hanzo_insights.client.batch_post")
+        cls.consumer_post_patcher = mock.patch("hanzo_insights.consumer.batch_post")
         cls.client_post_patcher.start()
         cls.consumer_post_patcher.start()
 
@@ -102,7 +102,7 @@ class TestFlagDefinitionCacheProvider(unittest.TestCase):
 class TestCacheInitialization(TestFlagDefinitionCacheProvider):
     """Tests for cache initialization behavior."""
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_uses_cached_data_when_should_fetch_returns_false(self, mock_get):
         """When should_fetch returns False and cache has data, use cached data."""
         self.cache_provider.should_fetch_return_value = False
@@ -124,7 +124,7 @@ class TestCacheInitialization(TestFlagDefinitionCacheProvider):
 
         client.join()
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_fetches_from_api_when_should_fetch_returns_true(self, mock_get):
         """When should_fetch returns True, fetch from API."""
         self.cache_provider.should_fetch_return_value = True
@@ -148,7 +148,7 @@ class TestCacheInitialization(TestFlagDefinitionCacheProvider):
 
         client.join()
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_emergency_fallback_when_cache_empty_and_no_flags(self, mock_get):
         """When should_fetch=False but cache is empty and no flags loaded, fetch anyway."""
         self.cache_provider.should_fetch_return_value = False
@@ -169,7 +169,7 @@ class TestCacheInitialization(TestFlagDefinitionCacheProvider):
 
         client.join()
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_preserves_existing_flags_when_cache_returns_none(self, mock_get):
         """When cache returns None but client has flags, preserve existing flags."""
         self.cache_provider.should_fetch_return_value = False
@@ -197,7 +197,7 @@ class TestCacheInitialization(TestFlagDefinitionCacheProvider):
 class TestFetchCoordination(TestFlagDefinitionCacheProvider):
     """Tests for fetch coordination between workers."""
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_calls_should_fetch_before_each_poll(self, mock_get):
         """should_fetch_flag_definitions is called before each poll cycle."""
         self.cache_provider.should_fetch_return_value = True
@@ -218,7 +218,7 @@ class TestFetchCoordination(TestFlagDefinitionCacheProvider):
 
         client.join()
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_does_not_call_on_received_when_fetch_skipped(self, mock_get):
         """on_flag_definitions_received is NOT called when fetch is skipped."""
         self.cache_provider.should_fetch_return_value = False
@@ -232,7 +232,7 @@ class TestFetchCoordination(TestFlagDefinitionCacheProvider):
 
         client.join()
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_stores_data_in_cache_after_api_fetch(self, mock_get):
         """on_flag_definitions_received receives the fetched data."""
         self.cache_provider.should_fetch_return_value = True
@@ -251,7 +251,7 @@ class TestFetchCoordination(TestFlagDefinitionCacheProvider):
 
         client.join()
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_304_not_modified_does_not_update_cache(self, mock_get):
         """When API returns 304 Not Modified, cache should not be updated."""
         self.cache_provider.should_fetch_return_value = True
@@ -293,7 +293,7 @@ class TestFetchCoordination(TestFlagDefinitionCacheProvider):
 class TestErrorHandling(TestFlagDefinitionCacheProvider):
     """Tests for error handling in cache provider operations."""
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_should_fetch_error_defaults_to_fetching(self, mock_get):
         """When should_fetch throws an error, default to fetching from API."""
         self.cache_provider.should_fetch_error = Exception("Lock acquisition failed")
@@ -313,7 +313,7 @@ class TestErrorHandling(TestFlagDefinitionCacheProvider):
 
         client.join()
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_get_error_falls_back_to_api_fetch(self, mock_get):
         """When get_flag_definitions throws an error, fetch from API."""
         self.cache_provider.should_fetch_return_value = False
@@ -331,7 +331,7 @@ class TestErrorHandling(TestFlagDefinitionCacheProvider):
 
         client.join()
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_on_received_error_keeps_flags_in_memory(self, mock_get):
         """When on_flag_definitions_received throws, flags are still in memory."""
         self.cache_provider.should_fetch_return_value = True
@@ -350,7 +350,7 @@ class TestErrorHandling(TestFlagDefinitionCacheProvider):
 
         client.join()
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_shutdown_error_is_logged_but_continues(self, mock_get):
         """When shutdown throws an error, it's logged but shutdown continues."""
         self.cache_provider.shutdown_error = Exception("Lock release failed")
@@ -372,7 +372,7 @@ class TestErrorHandling(TestFlagDefinitionCacheProvider):
 class TestShutdownLifecycle(TestFlagDefinitionCacheProvider):
     """Tests for shutdown lifecycle."""
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_shutdown_calls_cache_provider_shutdown(self, mock_get):
         """Client shutdown calls cache provider shutdown."""
         mock_get.return_value = GetResponse(
@@ -387,7 +387,7 @@ class TestShutdownLifecycle(TestFlagDefinitionCacheProvider):
 
         self.assertEqual(self.cache_provider.shutdown_call_count, 1)
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_shutdown_called_even_without_fetching(self, mock_get):
         """Shutdown is called even when cache was used instead of fetching."""
         self.cache_provider.should_fetch_return_value = False
@@ -400,7 +400,7 @@ class TestShutdownLifecycle(TestFlagDefinitionCacheProvider):
         # Shutdown should still be called
         self.assertEqual(self.cache_provider.shutdown_call_count, 1)
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_multiple_join_calls_only_shutdown_once(self, mock_get):
         """Calling join() multiple times should only call cache provider shutdown once."""
         mock_get.return_value = GetResponse(
@@ -423,7 +423,7 @@ class TestShutdownLifecycle(TestFlagDefinitionCacheProvider):
 class TestBackwardCompatibility(TestFlagDefinitionCacheProvider):
     """Tests for backward compatibility without cache provider."""
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_works_without_cache_provider(self, mock_get):
         """Client works normally without a cache provider configured."""
         mock_get.return_value = GetResponse(
@@ -451,7 +451,7 @@ class TestBackwardCompatibility(TestFlagDefinitionCacheProvider):
 class TestDataIntegrity(TestFlagDefinitionCacheProvider):
     """Tests for data integrity between cache and client state."""
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_cached_flags_available_for_evaluation(self, mock_get):
         """Flags loaded from cache are available for local evaluation."""
         self.cache_provider.should_fetch_return_value = False
@@ -483,7 +483,7 @@ class TestDataIntegrity(TestFlagDefinitionCacheProvider):
 
         client.join()
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_group_type_mapping_loaded_from_cache(self, mock_get):
         """Group type mapping is correctly loaded from cache."""
         self.cache_provider.should_fetch_return_value = False
@@ -497,7 +497,7 @@ class TestDataIntegrity(TestFlagDefinitionCacheProvider):
 
         client.join()
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_cohorts_loaded_from_cache(self, mock_get):
         """Cohorts are correctly loaded from cache."""
         self.cache_provider.should_fetch_return_value = False
@@ -510,7 +510,7 @@ class TestDataIntegrity(TestFlagDefinitionCacheProvider):
 
         client.join()
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_cache_updated_when_api_returns_new_data(self, mock_get):
         """State transition: cache has old data -> API returns new -> cache updated."""
         # Start with old cached data
@@ -556,7 +556,7 @@ class TestDataIntegrity(TestFlagDefinitionCacheProvider):
 class TestConcurrency(TestFlagDefinitionCacheProvider):
     """Tests for thread safety and concurrent access."""
 
-    @mock.patch("posthog.client.get")
+    @mock.patch("hanzo_insights.client.get")
     def test_concurrent_load_feature_flags_is_thread_safe(self, mock_get):
         """Multiple threads calling _load_feature_flags should not cause errors."""
         mock_get.return_value = GetResponse(

hanzo_insights/test/test_module.py

@@ -1,10 +1,10 @@
 import unittest
 
-from posthog import Posthog
+from hanzo_insights import Insights
 
 
 class TestModule(unittest.TestCase):
-    posthog = None
+    client = None
 
     def _assert_enqueue_result(self, result):
         self.assertEqual(type(result[0]), str)
@@ -14,19 +14,19 @@ class TestModule(unittest.TestCase):
 
     def setUp(self):
         self.failed = False
-        self.posthog = Posthog(
+        self.client = Insights(
             "testsecret", host="http://localhost:8000", on_error=self.failed
         )
 
     def test_track(self):
-        res = self.posthog.capture("python module event", distinct_id="distinct_id")
+        res = self.client.capture("python module event", distinct_id="distinct_id")
         self._assert_enqueue_result(res)
-        self.posthog.flush()
+        self.client.flush()
 
     def test_alias(self):
-        res = self.posthog.alias("previousId", "distinct_id")
+        res = self.client.alias("previousId", "distinct_id")
         self._assert_enqueue_result(res)
-        self.posthog.flush()
+        self.client.flush()
 
     def test_flush(self):
-        self.posthog.flush()
+        self.client.flush()

hanzo_insights/test/test_request.py

@@ -6,8 +6,8 @@ import mock
 import pytest
 import requests
 
-import posthog.request as request_module
-from posthog.request import (
+import hanzo_insights.request as request_module
+from hanzo_insights.request import (
     APIError,
     DatetimeSerializer,
     GetResponse,
@@ -23,7 +23,7 @@ from posthog.request import (
     get,
     set_socket_options,
 )
-from posthog.test.test_utils import TEST_API_KEY
+from hanzo_insights.test.test_utils import TEST_API_KEY
 
 
 @pytest.mark.parametrize(
@@ -128,7 +128,7 @@ class TestRequests(unittest.TestCase):
             }
         ).encode("utf-8")
 
-        with mock.patch("posthog.request._session.post", return_value=mock_response):
+        with mock.patch("hanzo_insights.request._session.post", return_value=mock_response):
             with self.assertRaises(QuotaLimitError) as cm:
                 decide("fake_key", "fake_host")
 
@@ -146,7 +146,7 @@ class TestRequests(unittest.TestCase):
             }
         ).encode("utf-8")
 
-        with mock.patch("posthog.request._session.post", return_value=mock_response):
+        with mock.patch("hanzo_insights.request._session.post", return_value=mock_response):
             response = decide("fake_key", "fake_host")
             self.assertEqual(response["featureFlags"], {"flag1": True})
 
@@ -154,7 +154,7 @@ class TestRequests(unittest.TestCase):
 class TestGet(unittest.TestCase):
     """Unit tests for the get() function HTTP-level behavior."""
 
-    @mock.patch("posthog.request._session.get")
+    @mock.patch("hanzo_insights.request._session.get")
     def test_get_returns_data_and_etag(self, mock_get):
         """Test that get() returns GetResponse with data and etag from headers."""
         mock_response = requests.Response()
@@ -172,7 +172,7 @@ class TestGet(unittest.TestCase):
         self.assertEqual(response.etag, '"abc123"')
         self.assertFalse(response.not_modified)
 
-    @mock.patch("posthog.request._session.get")
+    @mock.patch("hanzo_insights.request._session.get")
     def test_get_sends_if_none_match_header_when_etag_provided(self, mock_get):
         """Test that If-None-Match header is sent when etag parameter is provided."""
         mock_response = requests.Response()
@@ -186,7 +186,7 @@ class TestGet(unittest.TestCase):
         call_kwargs = mock_get.call_args[1]
         self.assertEqual(call_kwargs["headers"]["If-None-Match"], '"previous-etag"')
 
-    @mock.patch("posthog.request._session.get")
+    @mock.patch("hanzo_insights.request._session.get")
     def test_get_does_not_send_if_none_match_when_no_etag(self, mock_get):
         """Test that If-None-Match header is not sent when no etag provided."""
         mock_response = requests.Response()
@@ -199,7 +199,7 @@ class TestGet(unittest.TestCase):
         call_kwargs = mock_get.call_args[1]
         self.assertNotIn("If-None-Match", call_kwargs["headers"])
 
-    @mock.patch("posthog.request._session.get")
+    @mock.patch("hanzo_insights.request._session.get")
     def test_get_handles_304_not_modified(self, mock_get):
         """Test that 304 Not Modified response returns not_modified=True with no data."""
         mock_response = requests.Response()
@@ -216,7 +216,7 @@ class TestGet(unittest.TestCase):
         self.assertEqual(response.etag, '"unchanged-etag"')
         self.assertTrue(response.not_modified)
 
-    @mock.patch("posthog.request._session.get")
+    @mock.patch("hanzo_insights.request._session.get")
     def test_get_304_without_etag_header_uses_request_etag(self, mock_get):
         """Test that 304 response without ETag header falls back to request etag."""
         mock_response = requests.Response()
@@ -231,7 +231,7 @@ class TestGet(unittest.TestCase):
         self.assertTrue(response.not_modified)
         self.assertEqual(response.etag, '"original-etag"')
 
-    @mock.patch("posthog.request._session.get")
+    @mock.patch("hanzo_insights.request._session.get")
     def test_get_200_without_etag_header(self, mock_get):
         """Test that 200 response without ETag header returns None for etag."""
         mock_response = requests.Response()
@@ -246,7 +246,7 @@ class TestGet(unittest.TestCase):
         self.assertIsNone(response.etag)
         self.assertEqual(response.data, {"flags": []})
 
-    @mock.patch("posthog.request._session.get")
+    @mock.patch("hanzo_insights.request._session.get")
     def test_get_error_response_raises_api_error(self, mock_get):
         """Test that error responses raise APIError."""
         mock_response = requests.Response()
@@ -260,7 +260,7 @@ class TestGet(unittest.TestCase):
         self.assertEqual(ctx.exception.status, 401)
         self.assertEqual(ctx.exception.message, "Unauthorized")
 
-    @mock.patch("posthog.request._session.get")
+    @mock.patch("hanzo_insights.request._session.get")
     def test_get_sends_authorization_header(self, mock_get):
         """Test that Authorization header is sent with Bearer token."""
         mock_response = requests.Response()
@@ -273,7 +273,7 @@ class TestGet(unittest.TestCase):
         call_kwargs = mock_get.call_args[1]
         self.assertEqual(call_kwargs["headers"]["Authorization"], "Bearer my-api-key")
 
-    @mock.patch("posthog.request._session.get")
+    @mock.patch("hanzo_insights.request._session.get")
     def test_get_sends_user_agent_header(self, mock_get):
         """Test that User-Agent header is sent."""
         mock_response = requests.Response()
@@ -286,10 +286,10 @@ class TestGet(unittest.TestCase):
         call_kwargs = mock_get.call_args[1]
         self.assertIn("User-Agent", call_kwargs["headers"])
         self.assertTrue(
-            call_kwargs["headers"]["User-Agent"].startswith("posthog-python/")
+            call_kwargs["headers"]["User-Agent"].startswith("hanzo-insights-python/")
         )
 
-    @mock.patch("posthog.request._session.get")
+    @mock.patch("hanzo_insights.request._session.get")
     def test_get_passes_timeout(self, mock_get):
         """Test that timeout parameter is passed to the request."""
         mock_response = requests.Response()
@@ -302,7 +302,7 @@ class TestGet(unittest.TestCase):
         call_kwargs = mock_get.call_args[1]
         self.assertEqual(call_kwargs["timeout"], 30)
 
-    @mock.patch("posthog.request._session.get")
+    @mock.patch("hanzo_insights.request._session.get")
     def test_get_constructs_full_url(self, mock_get):
         """Test that host and url are combined correctly."""
         mock_response = requests.Response()
@@ -315,7 +315,7 @@ class TestGet(unittest.TestCase):
         call_args = mock_get.call_args[0]
         self.assertEqual(call_args[0], "https://example.com/api/flags")
 
-    @mock.patch("posthog.request._session.get")
+    @mock.patch("hanzo_insights.request._session.get")
     def test_get_removes_trailing_slash_from_host(self, mock_get):
         """Test that trailing slash is removed from host."""
         mock_response = requests.Response()
@@ -339,13 +339,13 @@ class TestGet(unittest.TestCase):
         ("https://us.posthog.com.rg.proxy.com", "https://us.posthog.com.rg.proxy.com"),
         ("app.posthog.com", "app.posthog.com"),
         ("eu.posthog.com", "eu.posthog.com"),
-        ("https://app.posthog.com", "https://us.i.posthog.com"),
-        ("https://eu.posthog.com", "https://eu.i.posthog.com"),
-        ("https://us.posthog.com", "https://us.i.posthog.com"),
-        ("https://app.posthog.com/", "https://us.i.posthog.com"),
-        ("https://eu.posthog.com/", "https://eu.i.posthog.com"),
-        ("https://us.posthog.com/", "https://us.i.posthog.com"),
-        (None, "https://us.i.posthog.com"),
+        ("https://app.posthog.com", "https://us.i.insights.hanzo.ai"),
+        ("https://eu.posthog.com", "https://eu.i.insights.hanzo.ai"),
+        ("https://us.posthog.com", "https://us.i.insights.hanzo.ai"),
+        ("https://app.posthog.com/", "https://us.i.insights.hanzo.ai"),
+        ("https://eu.posthog.com/", "https://eu.i.insights.hanzo.ai"),
+        ("https://us.posthog.com/", "https://us.i.insights.hanzo.ai"),
+        (None, "https://us.i.insights.hanzo.ai"),
     ],
 )
 def test_routing_to_custom_host(host, expected):
@@ -355,7 +355,7 @@ def test_routing_to_custom_host(host, expected):
 def test_enable_keep_alive_sets_socket_options():
     try:
         enable_keep_alive()
-        from posthog.request import _session
+        from hanzo_insights.request import _session
 
         adapter = _session.get_adapter("https://example.com")
         assert adapter.socket_options == KEEP_ALIVE_SOCKET_OPTIONS
@@ -367,7 +367,7 @@ def test_set_socket_options_clears_with_none():
     try:
         enable_keep_alive()
         set_socket_options(None)
-        from posthog.request import _session
+        from hanzo_insights.request import _session
 
         adapter = _session.get_adapter("https://example.com")
         assert adapter.socket_options is None
@@ -401,17 +401,17 @@ class TestFlagsSession(unittest.TestCase):
 
     def test_retry_status_forcelist_excludes_rate_limits(self):
         """Verify 429 (rate limit) is NOT retried - need to wait, not hammer."""
-        from posthog.request import RETRY_STATUS_FORCELIST
+        from hanzo_insights.request import RETRY_STATUS_FORCELIST
 
         self.assertNotIn(429, RETRY_STATUS_FORCELIST)
 
     def test_retry_status_forcelist_excludes_quota_errors(self):
         """Verify 402 (payment required/quota) is NOT retried - won't resolve."""
-        from posthog.request import RETRY_STATUS_FORCELIST
+        from hanzo_insights.request import RETRY_STATUS_FORCELIST
 
         self.assertNotIn(402, RETRY_STATUS_FORCELIST)
 
-    @mock.patch("posthog.request._get_flags_session")
+    @mock.patch("hanzo_insights.request._get_flags_session")
     def test_flags_uses_flags_session(self, mock_get_flags_session):
         """flags() uses the dedicated flags session, not the general session."""
         mock_response = requests.Response()
@@ -434,7 +434,7 @@ class TestFlagsSession(unittest.TestCase):
         mock_get_flags_session.assert_called_once()
         mock_session.post.assert_called_once()
 
-    @mock.patch("posthog.request._get_flags_session")
+    @mock.patch("hanzo_insights.request._get_flags_session")
     def test_flags_no_retry_on_quota_limit(self, mock_get_flags_session):
         """flags() raises QuotaLimitError without retrying (at application level)."""
         mock_response = requests.Response()
@@ -470,7 +470,7 @@ class TestFlagsSessionNetworkRetries(unittest.TestCase):
         retries on network-level failures (DNS failures, connection refused,
         connection reset, etc.) up to 2 times each.
         """
-        from posthog.request import _build_flags_session
+        from hanzo_insights.request import _build_flags_session
 
         session = _build_flags_session()
 
@@ -491,7 +491,7 @@ class TestFlagsSessionNetworkRetries(unittest.TestCase):
         This tests the status_forcelist configuration which specifies
         which HTTP status codes should trigger a retry.
         """
-        from posthog.request import _build_flags_session, RETRY_STATUS_FORCELIST
+        from hanzo_insights.request import _build_flags_session, RETRY_STATUS_FORCELIST
 
         session = _build_flags_session()
         adapter = session.get_adapter("https://test.posthog.com")
@@ -518,7 +518,7 @@ class TestFlagsSessionNetworkRetries(unittest.TestCase):
         """
         Verify that retries use exponential backoff to avoid thundering herd.
         """
-        from posthog.request import _build_flags_session
+        from hanzo_insights.request import _build_flags_session
 
         session = _build_flags_session()
         adapter = session.get_adapter("https://test.posthog.com")
@@ -545,7 +545,7 @@ class TestFlagsSessionRetryIntegration(unittest.TestCase):
         from http.server import HTTPServer, BaseHTTPRequestHandler
         from socketserver import ThreadingMixIn
         from urllib3.util.retry import Retry
-        from posthog.request import HTTPAdapterWithSocketOptions, RETRY_STATUS_FORCELIST
+        from hanzo_insights.request import HTTPAdapterWithSocketOptions, RETRY_STATUS_FORCELIST
 
         request_count = 0
 
@@ -631,7 +631,7 @@ class TestFlagsSessionRetryIntegration(unittest.TestCase):
         import socket
         import time
         from urllib3.util.retry import Retry
-        from posthog.request import HTTPAdapterWithSocketOptions, RETRY_STATUS_FORCELIST
+        from hanzo_insights.request import HTTPAdapterWithSocketOptions, RETRY_STATUS_FORCELIST
 
         # Get an available port by binding then closing a socket
         sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)

hanzo_insights/test/test_size_limited_dict.py

@@ -2,7 +2,7 @@ import unittest
 
 from parameterized import parameterized
 
-from posthog import utils
+from hanzo_insights import utils
 
 
 class TestSizeLimitedDict(unittest.TestCase):

hanzo_insights/test/test_types.py

@@ -2,7 +2,7 @@ import unittest
 
 from parameterized import parameterized
 
-from posthog.types import (
+from hanzo_insights.types import (
     FeatureFlag,
     FlagMetadata,
     FlagReason,

hanzo_insights/test/test_utils.py

@@ -13,8 +13,8 @@ from parameterized import parameterized
 from pydantic import BaseModel
 from pydantic.v1 import BaseModel as BaseModelV1
 
-from posthog import utils
-from posthog.types import FeatureFlagResult
+from hanzo_insights import utils
+from hanzo_insights.types import FeatureFlagResult
 
 TEST_API_KEY = "kOOlRy2QlMY9jHZQv0bKz0FZyazBUoY8Arj0lFVNjs4"
 FAKE_TEST_API_KEY = "random_key"
@@ -96,8 +96,8 @@ class TestUtils(unittest.TestCase):
 
     @parameterized.expand(
         [
-            ("http://posthog.io/", "http://posthog.io"),
-            ("http://posthog.io", "http://posthog.io"),
+            ("http://hanzo_insights.io/", "http://hanzo_insights.io"),
+            ("http://hanzo_insights.io", "http://hanzo_insights.io"),
             ("https://example.com/path/", "https://example.com/path"),
             ("https://example.com/path", "https://example.com/path"),
         ]

hanzo_insights/utils.py

@@ -16,7 +16,7 @@ import distro  # For Linux OS detection
 import six
 from dateutil.tz import tzlocal, tzutc
 
-log = logging.getLogger("posthog")
+log = logging.getLogger("hanzo_insights")
 
 
 def is_naive(dt):
@@ -277,7 +277,7 @@ class FlagCache:
 
 class RedisFlagCache:
     def __init__(
-        self, redis_client, default_ttl=300, stale_ttl=3600, key_prefix="posthog:flags:"
+        self, redis_client, default_ttl=300, stale_ttl=3600, key_prefix="insights:flags:"
     ):
         self.redis = redis_client
         self.default_ttl = default_ttl

hanzo_insights/version.py

@@ -0,0 +1 @@
+VERSION = "7.9.7"

insights/__init__.py

@@ -0,0 +1,3 @@
+# Convenience re-export so `from insights import Insights` works.
+from hanzo_insights import *  # noqa: F401, F403
+from hanzo_insights import Insights, Client  # noqa: F401

integration_tests/django5/test_exception_capture.py

@@ -1,10 +1,10 @@
 """
 Test that verifies exception capture functionality.
 
-These tests verify that exceptions are actually captured to PostHog, not just that
+These tests verify that exceptions are actually captured to Insights, not just that
 500 responses are returned.
 
-Without process_exception(), view exceptions are NOT captured to PostHog (v6.7.11 and earlier).
+Without process_exception(), view exceptions are NOT captured to Insights (v6.7.11 and earlier).
 With process_exception(), Django calls this method to capture exceptions before
 converting them to 500 responses.
 """
@@ -30,7 +30,7 @@ def asgi_app():
 @pytest.mark.asyncio
 async def test_async_exception_is_captured(asgi_app):
     """
-    Test that async view exceptions are captured to PostHog.
+    Test that async view exceptions are captured to Insights.
 
     The middleware's process_exception() method ensures exceptions are captured.
     Without it (v6.7.11 and earlier), exceptions are NOT captured even though 500 is returned.
@@ -50,8 +50,8 @@ async def test_async_exception_is_captured(asgi_app):
             }
         )
 
-    # Patch at the posthog module level where middleware imports from
-    with patch("posthog.capture_exception", side_effect=mock_capture):
+    # Patch at the hanzo_insights module level where middleware imports from
+    with patch("hanzo_insights.capture_exception", side_effect=mock_capture):
         async with AsyncClient(
             transport=ASGITransport(app=asgi_app), base_url="http://testserver"
         ) as ac:
@@ -60,8 +60,8 @@ async def test_async_exception_is_captured(asgi_app):
         # Django returns 500
         assert response.status_code == 500
 
-        # CRITICAL: Verify PostHog captured the exception
-        assert len(captured) > 0, "Exception was NOT captured to PostHog!"
+        # CRITICAL: Verify Insights captured the exception
+        assert len(captured) > 0, "Exception was NOT captured to Insights!"
 
         # Verify it's the right exception
         exception_data = captured[0]
@@ -72,7 +72,7 @@ async def test_async_exception_is_captured(asgi_app):
 @pytest.mark.asyncio
 async def test_sync_exception_is_captured(asgi_app):
     """
-    Test that sync view exceptions are captured to PostHog.
+    Test that sync view exceptions are captured to Insights.
 
     The middleware's process_exception() method ensures exceptions are captured.
     Without it (v6.7.11 and earlier), exceptions are NOT captured even though 500 is returned.
@@ -92,8 +92,8 @@ async def test_sync_exception_is_captured(asgi_app):
             }
         )
 
-    # Patch at the posthog module level where middleware imports from
-    with patch("posthog.capture_exception", side_effect=mock_capture):
+    # Patch at the hanzo_insights module level where middleware imports from
+    with patch("hanzo_insights.capture_exception", side_effect=mock_capture):
         async with AsyncClient(
             transport=ASGITransport(app=asgi_app), base_url="http://testserver"
         ) as ac:
@@ -102,8 +102,8 @@ async def test_sync_exception_is_captured(asgi_app):
         # Django returns 500
         assert response.status_code == 500
 
-        # CRITICAL: Verify PostHog captured the exception
-        assert len(captured) > 0, "Exception was NOT captured to PostHog!"
+        # CRITICAL: Verify Insights captured the exception
+        assert len(captured) > 0, "Exception was NOT captured to Insights!"
 
         # Verify it's the right exception
         exception_data = captured[0]

integration_tests/django5/test_middleware.py

@@ -1,5 +1,5 @@
 """
-Tests for PostHog Django middleware in async context.
+Tests for Insights Django middleware in async context.
 
 These tests verify that the middleware correctly handles:
 1. Async user access (request.auser() in Django 5)
@@ -103,7 +103,7 @@ async def test_async_authenticated_user_access(asgi_app):
 
     # Make request with session cookie - this should trigger the bug in v6.7.11
     # Disable exception capture to see the SynchronousOnlyOperation clearly
-    with override_settings(POSTHOG_MW_CAPTURE_EXCEPTIONS=False):
+    with override_settings(INSIGHTS_MW_CAPTURE_EXCEPTIONS=False):
         async with AsyncClient(
             transport=ASGITransport(app=asgi_app),
             base_url="http://testserver",
@@ -139,10 +139,10 @@ async def test_async_exception_capture(asgi_app):
     """
     Test that middleware handles exceptions from async views.
 
-    The middleware's process_exception() method captures view exceptions to PostHog
+    The middleware's process_exception() method captures view exceptions to Insights
     before Django converts them to 500 responses. This test verifies the exception
     causes a 500 response. See test_exception_capture.py for tests that verify
-    actual exception capture to PostHog.
+    actual exception capture to Insights.
     """
     async with AsyncClient(
         transport=ASGITransport(app=asgi_app), base_url="http://testserver"
@@ -158,7 +158,7 @@ async def test_sync_exception_capture(asgi_app):
     """
     Test that middleware handles exceptions from sync views.
 
-    The middleware's process_exception() method captures view exceptions to PostHog.
+    The middleware's process_exception() method captures view exceptions to Insights.
     This test verifies the exception causes a 500 response.
     """
     async with AsyncClient(

integration_tests/django5/testdjango/settings.py

@@ -47,7 +47,7 @@ MIDDLEWARE = [
     "django.contrib.auth.middleware.AuthenticationMiddleware",
     "django.contrib.messages.middleware.MessageMiddleware",
     "django.middleware.clickjacking.XFrameOptionsMiddleware",
-    "posthog.integrations.django.PosthogContextMiddleware",  # Test PostHog middleware
+    "hanzo_insights.integrations.django.InsightsContextMiddleware",  # Test Insights middleware
 ]
 
 ROOT_URLCONF = "testdjango.urls"
@@ -123,7 +123,7 @@ STATIC_URL = "static/"
 DEFAULT_AUTO_FIELD = "django.db.models.BigAutoField"
 
 
-# PostHog settings for testing
-POSTHOG_API_KEY = "test-key"
+# Insights settings for testing
+INSIGHTS_API_KEY = "test-key"
 POSTHOG_HOST = "https://app.posthog.com"
-POSTHOG_MW_CAPTURE_EXCEPTIONS = True
+INSIGHTS_MW_CAPTURE_EXCEPTIONS = True

integration_tests/django5/testdjango/views.py

@@ -1,5 +1,5 @@
 """
-Test views for validating PostHog middleware with Django 5 ASGI.
+Test views for validating Insights middleware with Django 5 ASGI.
 """
 
 from django.http import JsonResponse

mypy.ini

@@ -17,10 +17,10 @@ ignore_missing_imports = True
 [mypy-sentry_sdk.*]
 ignore_missing_imports = True
 
-[mypy-posthog.test.*]
+[mypy-hanzo_insights.test.*]
 ignore_errors = True
 
-[mypy-posthog.*.test.*]
+[mypy-hanzo_insights.*.test.*]
 ignore_errors = True
 
 [mypy-openai.*]

posthog/ai/utils.py

@@ -1,589 +0,0 @@
-import time
-import uuid
-from typing import Any, Callable, Dict, List, Optional, cast
-
-from posthog.client import Client as PostHogClient
-from posthog.ai.types import FormattedMessage, StreamingEventData, TokenUsage
-from posthog.ai.sanitization import (
-    sanitize_openai,
-    sanitize_anthropic,
-    sanitize_gemini,
-    sanitize_langchain,
-)
-
-
-def merge_usage_stats(
-    target: TokenUsage, source: TokenUsage, mode: str = "incremental"
-) -> None:
-    """
-    Merge streaming usage statistics into target dict, handling None values.
-
-    Supports two modes:
-    - "incremental": Add source values to target (for APIs that report new tokens)
-    - "cumulative": Replace target with source values (for APIs that report totals)
-
-    Args:
-        target: Dictionary to update with usage stats
-        source: TokenUsage that may contain None values
-        mode: Either "incremental" or "cumulative"
-    """
-    if mode == "incremental":
-        # Add new values to existing totals
-        source_input = source.get("input_tokens")
-        if source_input is not None:
-            current = target.get("input_tokens") or 0
-            target["input_tokens"] = current + source_input
-
-        source_output = source.get("output_tokens")
-        if source_output is not None:
-            current = target.get("output_tokens") or 0
-            target["output_tokens"] = current + source_output
-
-        source_cache_read = source.get("cache_read_input_tokens")
-        if source_cache_read is not None:
-            current = target.get("cache_read_input_tokens") or 0
-            target["cache_read_input_tokens"] = current + source_cache_read
-
-        source_cache_creation = source.get("cache_creation_input_tokens")
-        if source_cache_creation is not None:
-            current = target.get("cache_creation_input_tokens") or 0
-            target["cache_creation_input_tokens"] = current + source_cache_creation
-
-        source_reasoning = source.get("reasoning_tokens")
-        if source_reasoning is not None:
-            current = target.get("reasoning_tokens") or 0
-            target["reasoning_tokens"] = current + source_reasoning
-
-        source_web_search = source.get("web_search_count")
-        if source_web_search is not None:
-            current = target.get("web_search_count") or 0
-            target["web_search_count"] = max(current, source_web_search)
-
-    elif mode == "cumulative":
-        # Replace with latest values (already cumulative)
-        if source.get("input_tokens") is not None:
-            target["input_tokens"] = source["input_tokens"]
-        if source.get("output_tokens") is not None:
-            target["output_tokens"] = source["output_tokens"]
-        if source.get("cache_read_input_tokens") is not None:
-            target["cache_read_input_tokens"] = source["cache_read_input_tokens"]
-        if source.get("cache_creation_input_tokens") is not None:
-            target["cache_creation_input_tokens"] = source[
-                "cache_creation_input_tokens"
-            ]
-        if source.get("reasoning_tokens") is not None:
-            target["reasoning_tokens"] = source["reasoning_tokens"]
-        if source.get("web_search_count") is not None:
-            target["web_search_count"] = source["web_search_count"]
-
-    else:
-        raise ValueError(f"Invalid mode: {mode}. Must be 'incremental' or 'cumulative'")
-
-
-def get_model_params(kwargs: Dict[str, Any]) -> Dict[str, Any]:
-    """
-    Extracts model parameters from the kwargs dictionary.
-    """
-    model_params = {}
-    for param in [
-        "temperature",
-        "max_tokens",  # Deprecated field
-        "max_completion_tokens",
-        "top_p",
-        "frequency_penalty",
-        "presence_penalty",
-        "n",
-        "stop",
-        "stream",  # OpenAI-specific field
-        "streaming",  # Anthropic-specific field
-    ]:
-        if param in kwargs and kwargs[param] is not None:
-            model_params[param] = kwargs[param]
-    return model_params
-
-
-def get_usage(response, provider: str) -> TokenUsage:
-    """
-    Extract usage statistics from response based on provider.
-    Delegates to provider-specific converter functions.
-    """
-    if provider == "anthropic":
-        from posthog.ai.anthropic.anthropic_converter import (
-            extract_anthropic_usage_from_response,
-        )
-
-        return extract_anthropic_usage_from_response(response)
-    elif provider == "openai":
-        from posthog.ai.openai.openai_converter import (
-            extract_openai_usage_from_response,
-        )
-
-        return extract_openai_usage_from_response(response)
-    elif provider == "gemini":
-        from posthog.ai.gemini.gemini_converter import (
-            extract_gemini_usage_from_response,
-        )
-
-        return extract_gemini_usage_from_response(response)
-
-    return TokenUsage(input_tokens=0, output_tokens=0)
-
-
-def format_response(response, provider: str):
-    """
-    Format a regular (non-streaming) response.
-    """
-    if provider == "anthropic":
-        from posthog.ai.anthropic.anthropic_converter import format_anthropic_response
-
-        return format_anthropic_response(response)
-    elif provider == "openai":
-        from posthog.ai.openai.openai_converter import format_openai_response
-
-        return format_openai_response(response)
-    elif provider == "gemini":
-        from posthog.ai.gemini.gemini_converter import format_gemini_response
-
-        return format_gemini_response(response)
-    return []
-
-
-def extract_available_tool_calls(provider: str, kwargs: Dict[str, Any]):
-    """
-    Extract available tool calls for the given provider.
-    """
-    if provider == "anthropic":
-        from posthog.ai.anthropic.anthropic_converter import extract_anthropic_tools
-
-        return extract_anthropic_tools(kwargs)
-    elif provider == "gemini":
-        from posthog.ai.gemini.gemini_converter import extract_gemini_tools
-
-        return extract_gemini_tools(kwargs)
-    elif provider == "openai":
-        from posthog.ai.openai.openai_converter import extract_openai_tools
-
-        return extract_openai_tools(kwargs)
-    return None
-
-
-def merge_system_prompt(
-    kwargs: Dict[str, Any], provider: str
-) -> List[FormattedMessage]:
-    """
-    Merge system prompts and format messages for the given provider.
-    """
-    if provider == "anthropic":
-        from posthog.ai.anthropic.anthropic_converter import format_anthropic_input
-
-        messages = kwargs.get("messages") or []
-        system = kwargs.get("system")
-        return format_anthropic_input(messages, system)
-    elif provider == "gemini":
-        from posthog.ai.gemini.gemini_converter import format_gemini_input_with_system
-
-        contents = kwargs.get("contents", [])
-        config = kwargs.get("config")
-        return format_gemini_input_with_system(contents, config)
-    elif provider == "openai":
-        from posthog.ai.openai.openai_converter import format_openai_input
-
-        # For OpenAI, handle both Chat Completions and Responses API
-        messages_param = kwargs.get("messages")
-        input_param = kwargs.get("input")
-
-        # Get base formatted messages
-        messages = format_openai_input(messages_param, input_param)
-
-        # Check if system prompt is provided as a separate parameter
-        if kwargs.get("system") is not None:
-            has_system = any(msg.get("role") == "system" for msg in messages)
-            if not has_system:
-                system_msg = cast(
-                    FormattedMessage,
-                    {"role": "system", "content": kwargs.get("system")},
-                )
-                messages = [system_msg] + messages
-
-        # For Responses API, add instructions to the system prompt if provided
-        if kwargs.get("instructions") is not None:
-            # Find the system message if it exists
-            system_idx = next(
-                (i for i, msg in enumerate(messages) if msg.get("role") == "system"),
-                None,
-            )
-
-            if system_idx is not None:
-                # Append instructions to existing system message
-                system_content = messages[system_idx].get("content", "")
-                messages[system_idx]["content"] = (
-                    f"{system_content}\n\n{kwargs.get('instructions')}"
-                )
-            else:
-                # Create a new system message with instructions
-                instruction_msg = cast(
-                    FormattedMessage,
-                    {"role": "system", "content": kwargs.get("instructions")},
-                )
-                messages = [instruction_msg] + messages
-
-        return messages
-
-    # Default case - return empty list
-    return []
-
-
-def call_llm_and_track_usage(
-    posthog_distinct_id: Optional[str],
-    ph_client: PostHogClient,
-    provider: str,
-    posthog_trace_id: Optional[str],
-    posthog_properties: Optional[Dict[str, Any]],
-    posthog_privacy_mode: bool,
-    posthog_groups: Optional[Dict[str, Any]],
-    base_url: str,
-    call_method: Callable[..., Any],
-    **kwargs: Any,
-) -> Any:
-    """
-    Common usage-tracking logic for both sync and async calls.
-    call_method: the llm call method (e.g. openai.chat.completions.create)
-    """
-    start_time = time.time()
-    response = None
-    error = None
-    http_status = 200
-    usage: TokenUsage = TokenUsage()
-    error_params: Dict[str, Any] = {}
-
-    try:
-        response = call_method(**kwargs)
-    except Exception as exc:
-        error = exc
-        http_status = getattr(
-            exc, "status_code", 0
-        )  # default to 0 becuase its likely an SDK error
-        error_params = {
-            "$ai_is_error": True,
-            "$ai_error": exc.__str__(),
-        }
-    finally:
-        end_time = time.time()
-        latency = end_time - start_time
-
-        if posthog_trace_id is None:
-            posthog_trace_id = str(uuid.uuid4())
-
-        if response and (
-            hasattr(response, "usage")
-            or (provider == "gemini" and hasattr(response, "usage_metadata"))
-        ):
-            usage = get_usage(response, provider)
-
-        messages = merge_system_prompt(kwargs, provider)
-        sanitized_messages = sanitize_messages(messages, provider)
-
-        event_properties = {
-            "$ai_provider": provider,
-            "$ai_model": kwargs.get("model") or getattr(response, "model", None),
-            "$ai_model_parameters": get_model_params(kwargs),
-            "$ai_input": with_privacy_mode(
-                ph_client, posthog_privacy_mode, sanitized_messages
-            ),
-            "$ai_output_choices": with_privacy_mode(
-                ph_client, posthog_privacy_mode, format_response(response, provider)
-            ),
-            "$ai_http_status": http_status,
-            "$ai_input_tokens": usage.get("input_tokens", 0),
-            "$ai_output_tokens": usage.get("output_tokens", 0),
-            "$ai_latency": latency,
-            "$ai_trace_id": posthog_trace_id,
-            "$ai_base_url": str(base_url),
-            **(posthog_properties or {}),
-            **(error_params or {}),
-        }
-
-        available_tool_calls = extract_available_tool_calls(provider, kwargs)
-
-        if available_tool_calls:
-            event_properties["$ai_tools"] = available_tool_calls
-
-        cache_read = usage.get("cache_read_input_tokens")
-        if cache_read is not None and cache_read > 0:
-            event_properties["$ai_cache_read_input_tokens"] = cache_read
-
-        cache_creation = usage.get("cache_creation_input_tokens")
-        if cache_creation is not None and cache_creation > 0:
-            event_properties["$ai_cache_creation_input_tokens"] = cache_creation
-
-        reasoning = usage.get("reasoning_tokens")
-        if reasoning is not None and reasoning > 0:
-            event_properties["$ai_reasoning_tokens"] = reasoning
-
-        web_search_count = usage.get("web_search_count")
-        if web_search_count is not None and web_search_count > 0:
-            event_properties["$ai_web_search_count"] = web_search_count
-
-        if posthog_distinct_id is None:
-            event_properties["$process_person_profile"] = False
-
-        # Process instructions for Responses API
-        if provider == "openai" and kwargs.get("instructions") is not None:
-            event_properties["$ai_instructions"] = with_privacy_mode(
-                ph_client, posthog_privacy_mode, kwargs.get("instructions")
-            )
-
-        # send the event to posthog
-        if hasattr(ph_client, "capture") and callable(ph_client.capture):
-            ph_client.capture(
-                distinct_id=posthog_distinct_id or posthog_trace_id,
-                event="$ai_generation",
-                properties=event_properties,
-                groups=posthog_groups,
-            )
-
-    if error:
-        raise error
-
-    return response
-
-
-async def call_llm_and_track_usage_async(
-    posthog_distinct_id: Optional[str],
-    ph_client: PostHogClient,
-    provider: str,
-    posthog_trace_id: Optional[str],
-    posthog_properties: Optional[Dict[str, Any]],
-    posthog_privacy_mode: bool,
-    posthog_groups: Optional[Dict[str, Any]],
-    base_url: str,
-    call_async_method: Callable[..., Any],
-    **kwargs: Any,
-) -> Any:
-    start_time = time.time()
-    response = None
-    error = None
-    http_status = 200
-    usage: TokenUsage = TokenUsage()
-    error_params: Dict[str, Any] = {}
-
-    try:
-        response = await call_async_method(**kwargs)
-    except Exception as exc:
-        error = exc
-        http_status = getattr(
-            exc, "status_code", 0
-        )  # default to 0 because its likely an SDK error
-        error_params = {
-            "$ai_is_error": True,
-            "$ai_error": exc.__str__(),
-        }
-    finally:
-        end_time = time.time()
-        latency = end_time - start_time
-
-        if posthog_trace_id is None:
-            posthog_trace_id = str(uuid.uuid4())
-
-        if response and (
-            hasattr(response, "usage")
-            or (provider == "gemini" and hasattr(response, "usage_metadata"))
-        ):
-            usage = get_usage(response, provider)
-
-        messages = merge_system_prompt(kwargs, provider)
-        sanitized_messages = sanitize_messages(messages, provider)
-
-        event_properties = {
-            "$ai_provider": provider,
-            "$ai_model": kwargs.get("model") or getattr(response, "model", None),
-            "$ai_model_parameters": get_model_params(kwargs),
-            "$ai_input": with_privacy_mode(
-                ph_client, posthog_privacy_mode, sanitized_messages
-            ),
-            "$ai_output_choices": with_privacy_mode(
-                ph_client, posthog_privacy_mode, format_response(response, provider)
-            ),
-            "$ai_http_status": http_status,
-            "$ai_input_tokens": usage.get("input_tokens", 0),
-            "$ai_output_tokens": usage.get("output_tokens", 0),
-            "$ai_latency": latency,
-            "$ai_trace_id": posthog_trace_id,
-            "$ai_base_url": str(base_url),
-            **(posthog_properties or {}),
-            **(error_params or {}),
-        }
-
-        available_tool_calls = extract_available_tool_calls(provider, kwargs)
-
-        if available_tool_calls:
-            event_properties["$ai_tools"] = available_tool_calls
-
-        cache_read = usage.get("cache_read_input_tokens")
-        if cache_read is not None and cache_read > 0:
-            event_properties["$ai_cache_read_input_tokens"] = cache_read
-
-        cache_creation = usage.get("cache_creation_input_tokens")
-        if cache_creation is not None and cache_creation > 0:
-            event_properties["$ai_cache_creation_input_tokens"] = cache_creation
-
-        reasoning = usage.get("reasoning_tokens")
-        if reasoning is not None and reasoning > 0:
-            event_properties["$ai_reasoning_tokens"] = reasoning
-
-        web_search_count = usage.get("web_search_count")
-        if web_search_count is not None and web_search_count > 0:
-            event_properties["$ai_web_search_count"] = web_search_count
-
-        if posthog_distinct_id is None:
-            event_properties["$process_person_profile"] = False
-
-        # Process instructions for Responses API
-        if provider == "openai" and kwargs.get("instructions") is not None:
-            event_properties["$ai_instructions"] = with_privacy_mode(
-                ph_client, posthog_privacy_mode, kwargs.get("instructions")
-            )
-
-        # send the event to posthog
-        if hasattr(ph_client, "capture") and callable(ph_client.capture):
-            ph_client.capture(
-                distinct_id=posthog_distinct_id or posthog_trace_id,
-                event="$ai_generation",
-                properties=event_properties,
-                groups=posthog_groups,
-            )
-
-    if error:
-        raise error
-
-    return response
-
-
-def sanitize_messages(data: Any, provider: str) -> Any:
-    """Sanitize messages using provider-specific sanitization functions."""
-    if provider == "anthropic":
-        return sanitize_anthropic(data)
-    elif provider == "openai":
-        return sanitize_openai(data)
-    elif provider == "gemini":
-        return sanitize_gemini(data)
-    elif provider == "langchain":
-        return sanitize_langchain(data)
-    return data
-
-
-def with_privacy_mode(ph_client: PostHogClient, privacy_mode: bool, value: Any):
-    if ph_client.privacy_mode or privacy_mode:
-        return None
-    return value
-
-
-def capture_streaming_event(
-    ph_client: PostHogClient,
-    event_data: StreamingEventData,
-):
-    """
-    Unified streaming event capture for all LLM providers.
-
-    This function handles the common logic for capturing streaming events across all providers.
-    All provider-specific formatting should be done BEFORE calling this function.
-
-    The function handles:
-    - Building PostHog event properties
-    - Extracting and adding tools based on provider
-    - Applying privacy mode
-    - Adding special token fields (cache, reasoning)
-    - Provider-specific fields (e.g., OpenAI instructions)
-    - Sending the event to PostHog
-
-    Args:
-        ph_client: PostHog client instance
-        event_data: Standardized streaming event data containing all necessary information
-    """
-    trace_id = event_data.get("trace_id") or str(uuid.uuid4())
-
-    # Build base event properties
-    event_properties = {
-        "$ai_provider": event_data["provider"],
-        "$ai_model": event_data["model"],
-        "$ai_model_parameters": get_model_params(event_data["kwargs"]),
-        "$ai_input": with_privacy_mode(
-            ph_client,
-            event_data["privacy_mode"],
-            event_data["formatted_input"],
-        ),
-        "$ai_output_choices": with_privacy_mode(
-            ph_client,
-            event_data["privacy_mode"],
-            event_data["formatted_output"],
-        ),
-        "$ai_http_status": 200,
-        "$ai_input_tokens": event_data["usage_stats"].get("input_tokens", 0),
-        "$ai_output_tokens": event_data["usage_stats"].get("output_tokens", 0),
-        "$ai_latency": event_data["latency"],
-        "$ai_trace_id": trace_id,
-        "$ai_base_url": str(event_data["base_url"]),
-        **(event_data.get("properties") or {}),
-    }
-
-    # Extract and add tools based on provider
-    available_tools = extract_available_tool_calls(
-        event_data["provider"],
-        event_data["kwargs"],
-    )
-    if available_tools:
-        event_properties["$ai_tools"] = available_tools
-
-    # Add optional token fields
-    # For Anthropic, always include cache fields even if 0 (backward compatibility)
-    # For others, only include if present and non-zero
-    if event_data["provider"] == "anthropic":
-        # Anthropic always includes cache fields
-        cache_read = event_data["usage_stats"].get("cache_read_input_tokens", 0)
-        cache_creation = event_data["usage_stats"].get("cache_creation_input_tokens", 0)
-        event_properties["$ai_cache_read_input_tokens"] = cache_read
-        event_properties["$ai_cache_creation_input_tokens"] = cache_creation
-    else:
-        # Other providers only include if non-zero
-        optional_token_fields = [
-            "cache_read_input_tokens",
-            "cache_creation_input_tokens",
-            "reasoning_tokens",
-        ]
-
-        for field in optional_token_fields:
-            value = event_data["usage_stats"].get(field)
-            if value is not None and isinstance(value, int) and value > 0:
-                event_properties[f"$ai_{field}"] = value
-
-    # Add web search count if present (all providers)
-    web_search_count = event_data["usage_stats"].get("web_search_count")
-    if (
-        web_search_count is not None
-        and isinstance(web_search_count, int)
-        and web_search_count > 0
-    ):
-        event_properties["$ai_web_search_count"] = web_search_count
-
-    # Handle provider-specific fields
-    if (
-        event_data["provider"] == "openai"
-        and event_data["kwargs"].get("instructions") is not None
-    ):
-        event_properties["$ai_instructions"] = with_privacy_mode(
-            ph_client,
-            event_data["privacy_mode"],
-            event_data["kwargs"]["instructions"],
-        )
-
-    if event_data.get("distinct_id") is None:
-        event_properties["$process_person_profile"] = False
-
-    # Send event to PostHog
-    if hasattr(ph_client, "capture"):
-        ph_client.capture(
-            distinct_id=event_data.get("distinct_id") or trace_id,
-            event="$ai_generation",
-            properties=event_properties,
-            groups=event_data.get("groups"),
-        )