GitTLDR transforms how development teams understand and collaborate on code repositories. Built from frustration with lengthy code reviews and scattered repository insights, this platform combines AI-powered analysis with real-time team coordination to make repository management actually efficient.
Extract meaningful insights from your repositories without drowning in documentation. Our AI analyzes commit patterns, identifies code relationships, and generates summaries that actually help developers understand what's happening.
Transform GitHub issues into working code automatically. Our AI analyzes issue descriptions, understands your codebase context through vector embeddings, generates validated fixes using Tree-of-Thought reasoning, and creates pull requests with confidence scoring.
Ask questions about your codebase in natural language. Get answers with citations to specific files and lines, powered by RAG (Retrieval-Augmented Generation) and Google Gemini AI.
Connect your development meetings to repository context. Record team discussions, automatically transcribe them, and link decisions back to specific code changes or repository events.
See who's working on what, track changes as they happen, and get notified about relevant updates.
GitTLDR/
├── frontend/ # Next.js 15 dashboard and web interface (port 3000)
├── node-worker/ # Express.js API server and task orchestration (port 3001)
├── python-worker/ # FastAPI AI processing and embeddings (port 8001)
└── load-tests/ # k6 performance testing scripts
- Frontend: User interface, authentication, real-time updates via WebSocket
- Node Worker: GitHub integration, webhooks, queue management, database operations
- Python Worker: AI summarization, vector embeddings, semantic search, issue fix generation
- Automatic commit history analysis and pattern detection
- Code relationship mapping across files and contributors
- Integration health monitoring and dependency tracking
- Intelligent code and documentation summarization using Google Gemini
- Vector-based semantic search across repositories and meeting transcripts
- Context-aware explanations for complex code changes
- Automated issue analysis and fix generation
- Real-time notifications for relevant repository events
- Meeting transcription with automatic linking to repository context
- Shared workspace for repository insights and team coordination
- AI-powered analysis of GitHub issue descriptions
- Precision code retrieval using vector embeddings and AST parsing
- Multi-layer validation with confidence scoring
- Automatic PR creation with detailed change explanations
| Component | Technology | Why We Chose It |
|---|---|---|
| Frontend | Next.js 15, React 19, TypeScript | Server-side rendering, excellent TypeScript support |
| API Layer | Node.js, Express, Prisma | Familiar ecosystem, great GitHub API integration |
| AI Processing | Python, FastAPI, Google Gemini | Best-in-class ML libraries, async processing |
| Database | PostgreSQL, Redis, Qdrant | ACID transactions, fast caching, vector search |
| Knowledge Graph | Neo4j | Code relationship mapping |
| Infrastructure | Docker, Render, Vercel | Simple deployment, automatic scaling |
- Node.js 18+ and Python 3.11+
- PostgreSQL and Redis instances
- GitHub OAuth application
- Google Gemini API key
-
Clone and configure:
git clone https://github.com/MugundhanY/GitTLDR.git cd GitTLDR cp frontend/.env.example frontend/.env cp node-worker/.env.example node-worker/.env cp python-worker/.env.example python-worker/.env -
Install dependencies:
cd frontend && npm install && cd .. cd node-worker && npm install && cd .. cd python-worker && pip install -r requirements.txt && cd ..
-
Set up database:
cd frontend npx prisma migrate dev npx prisma generate cd ..
-
Start all services (in separate terminals):
# Frontend (http://localhost:3000) cd frontend && npm run dev # Node worker (http://localhost:3001) cd node-worker && npm run dev # Python worker (http://localhost:8001) # Option 1: Combined mode (API + background worker) cd python-worker && python main.py # Option 2: Separate processes cd python-worker && python api_server.py # API only cd python-worker && python worker.py # Worker only
GitHub Integration
GITHUB_CLIENT_ID=your_github_app_client_id
GITHUB_CLIENT_SECRET=your_github_app_secret
AI Services
GEMINI_API_KEY=your_gemini_key
GEMINI_API_KEYS=key1,key2,key3 # Multiple keys for rotation
Infrastructure
DATABASE_URL=postgresql://user:password@localhost:5432/gittldr
REDIS_URL=redis://localhost:6379
QDRANT_URL=http://localhost:6333
Independent scaling of AI processing (CPU/GPU intensive) versus web traffic (memory/network intensive). Node.js handles GitHub API rate limits, while Python excels at AI/ML workloads.
Repository analysis can take minutes for large repos. Redis lists and pub/sub process work asynchronously with real-time progress updates.
Semantic search using Qdrant that understands context and relationships, not just keywords.
# Python Worker (25 tests)
cd python-worker && pip install -r requirements-test.txt && pytest -v
# Node Worker (13 tests)
cd node-worker && npm install && npm testTest results:
# Install k6: https://k6.io/docs/get-started/installation/
cd load-tests
# Quick test (services must be running)
k6 run --vus 2 --duration 10s k6-health.js
# Test deployed services
k6 run -e PYTHON_URL=https://your-app.onrender.com -e NODE_URL=https://your-node.onrender.com k6-health.jsGitHub Actions runs on every push:
- Python tests (pytest)
- Node tests (Jest)
- Frontend build verification
- Deployment to Render/Vercel on main branch
Recommended: Render + Vercel (Free Tier)
- Backend services on Render using
render.yamlBlueprint - Frontend on Vercel (global CDN)
- Neon for PostgreSQL, Upstash for Redis, Qdrant Cloud for vectors
See DEPLOYMENT.md for complete guide.
Alternative: Railway
- $5 free credit, supports background workers
- Better for full functionality
Documentation: Each service directory contains detailed setup instructions, API documentation, and deployment guides.