extension/docs/COMPARISON.md

Dev vs Traditional AI Tools - Technical Comparison

Architecture Differences

Traditional AI Tools

• Sequential Processing: One query at a time
• Limited Context: 8K-128K token windows
• Session-based Memory: Context resets between sessions
• Single Model: Usually locked to one AI provider
• File-by-file Operations: Manual coordination needed

Dev Platform

• Parallel Processing: Unlimited concurrent agents
• Full Codebase Context: Entire repository indexed
• Persistent Memory: Context maintained across sessions
• Multi-Model: 200+ models via Hanzo AI gateway
• Coordinated Operations: Automatic cross-file changes

---

Performance Comparison

Task: Add Authentication System

Traditional Approach

1. Query AI for auth model (2 min)
2. Debug imports, ask again (5 min)
3. Query for middleware (2 min)
4. Manually integrate (10 min)
5. Query for endpoints (2 min)
6. Fix inconsistencies (15 min)
7. Query for tests (2 min)
8. Write missing tests (20 min)
Total: ~60 minutes

Dev Approach

dev enhance "add JWT auth with refresh tokens"
- Spawns 5 parallel agents
- All work simultaneously
- Automatic integration
Total: ~1 minute

---

Feature Matrix

Feature	Traditional AI	Dev Platform
Parallel Execution	❌ No	✅ Unlimited agents
Context Window	8K-128K tokens	Entire codebase
Memory Persistence	Per session	Forever
Multi-file Coordination	Manual	Automatic
Git Integration	Basic	Full worktree support
Import Verification	No	Yes
Model Options	1-5	200+
Local LLM Support	Varies	Built-in
Team Collaboration	No	Yes
Cloud Integration	Limited	Full via cloud.hanzo.ai

---

Speed Multipliers

Why Dev is Faster

1. Parallel vs Sequential

   • Traditional: Each step waits for previous
   • Dev: All steps execute simultaneously
   • Speed gain: 5-10x minimum

2. Context Availability

   • Traditional: Re-explain codebase each query
   • Dev: Instant access to all code
   • Speed gain: 2-3x

3. Accuracy

   • Traditional: Debug hallucinated code
   • Dev: Verified imports only
   • Speed gain: 2-4x

4. Coordination

   • Traditional: Manual file updates
   • Dev: Automatic cross-file changes
   • Speed gain: 3-5x

Combined Effect: 30x-100x faster for complex tasks

---

Integration Comparison

Traditional Tools

• Separate logins for each service
• Multiple API keys to manage
• No unified billing
• Context not shared between tools

Dev Platform

• Single login via cloud.hanzo.ai
• One API key for everything
• Unified billing and usage
• Shared context across all tools

---

Cost Structure

Traditional AI Tools

• Pay per tool (Claude, GPT-4, etc.)
• Pay for wasted tokens on context
• Pay in developer time for debugging
• Hidden costs in context switching

Dev Platform

• Single subscription
• Efficient token usage
• Minimal debugging needed
• Includes all tools and models

---

Get Started

Experience the difference yourself:

npm install -g @hanzo/dev
dev login
dev enhance "your first feature"

Sign up at cloud.hanzo.ai →