Analyze your AI coding assistant conversations. Track what works, find patterns, improve your workflow.
# Clone and setup
git clone https://github.com/kulesh/catsyphon.git && cd catsyphon
cp .env.example .env
# Start everything
docker-compose up -d
cd backend && uv sync --all-extras && uv run alembic upgrade head
uv run catsyphon serve &
cd ../frontend && pnpm install && pnpm devOpen http://localhost:5173 and start ingesting your Claude Code logs.
Where are my logs? ~/.claude/projects/ contains your Claude Code conversation history.
Prerequisites: Python 3.11+, Node 20+, Docker, uv, pnpm
Use mise to install tools automatically:
mise installOptional: Configure
LLM_PROVIDER,LLM_MODEL, and the matching provider API key in.envfor AI-powered analytics/tagging.
See sessions, activity, and navigate to detailed analytics.
Search and filter through all your AI-assisted coding sessions. See session names, git branches, token usage, and plan indicators at a glance.
Point at your log directories and watch them auto-import. Supports live sessions.
See Claude's planning process - view plans, track iterations, and compare changes between versions with diff view.
Get an effectiveness score based on success rates, session patterns, and outcomes.
See what's working, what's not, and actionable recommendations.
Track sentiment over time, see which tools are used most, identify top problems.
Collaboration quality, agent effectiveness, and scope clarity scores with weekly trends.
Discover your workflow patterns, learning opportunities, and anti-patterns.
See which files changed most, lines added/deleted across all sessions.
| Agent | Status |
|---|---|
| Claude Code | Built-in |
| OpenAI Codex | Built-in |
| Others | Plugin SDK |
Backend: Python 3.11, FastAPI, PostgreSQL, SQLAlchemy, provider-agnostic LLM adapters (OpenAI/Anthropic/Google) Frontend: React 19, TypeScript, Vite, Tailwind, shadcn/ui
# Quick start - runs full stack (Colima, PostgreSQL, API, Frontend)
./scripts/dev.sh start
# Or start components separately
./scripts/dev.sh backend # Backend only
./scripts/dev.sh frontend # Frontend only
# Other commands
./scripts/dev.sh status # Check all services
./scripts/dev.sh stop # Stop everything
./scripts/dev.sh logs # Stream logs# Backend tests and quality
cd backend
uv run pytest # Tests
uv run mypy src/ # Type check
uv run black src/ tests/ # Format
# Frontend tests and quality
cd frontend
pnpm test # Tests
pnpm tsc --noEmit # Type check| I want to... | Start here |
|---|---|
| Analyze my own logs locally | Getting Started |
| Deploy for my team/company | Enterprise Deployment |
| Run multi-Mac edge sensors into one central collector | Edge Sensors (macOS) |
| Build a parser plugin | Parser SDK |
| Build a collector client | Collector SDK |
| Contribute to CatSyphon | Contributing |
See the Documentation Index for the full navigation hub.
- Getting Started - Quick start for individual users
- Enterprise Deployment - Team/organization setup
- Parser SDK - Add support for other AI assistants
- Collector SDK - Build HTTP clients for CatSyphon
- Collector Protocol - Event streaming API specification
- Edge Sensors (macOS) - 24/7 multi-machine setup to central collector
- OTEL Ingestion - Ingest Codex OTLP events
- API Reference - REST API documentation
- Architecture - System design and diagrams
- Contributing - Development workflow
MIT