Intelligent developer assistant that extends Claude's context through MCP+RAG, providing seamless access to codebase knowledge while optimizing costs through smart model routing and caching.
Signal Hub is part of the Signal ecosystem, serving as the central convergence point where all your code context comes together for intelligent AI routing. It acts as an MCP (Model Context Protocol) server that gives Claude unlimited effective context through RAG (Retrieval-Augmented Generation) while intelligently routing queries to the most cost-effective model.
The community edition provides core functionality for extending Claude's context:
- ✅ MCP server integration with Claude Code
- ✅ Codebase indexing and semantic search
- ✅ Basic RAG retrieval and context assembly
- ✅ Simple rule-based model routing
- ✅ Semantic caching for repeated queries
- ✅ Manual escalation to premium models
- ✅ Basic cost tracking
Advanced features for teams and power users:
- 🚀 ML-powered intelligent model routing
- 📈 Learning algorithms that improve over time
- 💰 Advanced cost optimization (save 70%+ on AI costs)
- 📊 Detailed analytics and ROI reporting
- 👥 Team collaboration and sharing
- 🔐 Enterprise security and compliance
- ⚡ Priority support and SLAs
Early Access Program: Currently, all features are available to early adopters to help us refine the product and pricing. Join our Discord to participate!
- 🚀 Unlimited Context: Access your entire codebase through semantic search
- 💰 Cost Optimization: Automatic routing between Haiku, Sonnet, and Opus based on complexity
- ⚡ Lightning Fast: Semantic caching for instant responses to similar queries
- 🔄 Continuous Learning: Improves routing decisions based on user feedback (Pro)
- 🛠️ Zero Friction: Works out of the box with minimal configuration
- 🌍 Open Source: Community-driven development with transparent roadmap
- Python 3.11 or higher
- Docker Desktop
- Git
git clone https://github.com/wespiper/signal-hub.git
cd signal-hub
./scripts/setup.shgit clone https://github.com/wespiper/signal-hub.git
cd signal-hub
.\scripts\setup.ps1- Install Poetry:
curl -sSL https://install.python-poetry.org | python3 -- Install dependencies:
poetry install- Configure Signal Hub:
cp config/.env.example config/.env
# Add your OpenAI API key for embeddings- Start services:
cd docker && docker compose up -d
cd .. && signal-hub serve --config config/dev.yaml- Connect with Claude Code by adding Signal Hub to your MCP servers configuration.
Signal Hub provides several MCP tools for intelligent code navigation:
search_code: Semantic search across your entire codebaseexplain_code: Get detailed explanations with relevant contextfind_similar: Find code similar to a given snippetget_context: Get relevant context for your current taskescalate_query: Manual escalation to better model (coming in Sprint 3)
Signal Hub consists of several key components:
- MCP Server: Handles communication with Claude Code
- Indexing Pipeline: Scans and embeds your codebase
- RAG System: Retrieves relevant context for queries
- Model Router: Intelligently selects the appropriate model
- Cache Layer: Stores and retrieves similar queries
# Run all tests with coverage
make test
# Run specific test categories
pytest tests/unit/
pytest tests/integration/
# Generate coverage report
make test-coveragemake format # Format code with Black
make lint # Run Ruff linter
make type-check # Type checking with MyPy
make quality # Run all quality checks-
Start Development Environment:
make dev
-
Run with Early Access Features:
SIGNAL_HUB_EARLY_ACCESS=true signal-hub serve
-
Monitor Logs:
# JSON logs for production SIGNAL_HUB_LOG_FORMAT=json signal-hub serve # Check metrics endpoint curl http://localhost:3334/metrics
We welcome contributions! Please see our Contributing Guidelines for details.
- 🐛 Report bugs using our issue template
- 💡 Request features using our feature template
- 🔧 Submit PRs following our PR template
- ✅ Basic MCP server implementation with plugin architecture
- ✅ Codebase indexing and embedding pipeline
- ✅ File parser framework (Python, JavaScript, Markdown)
- ✅ ChromaDB integration for vector storage
- ✅ Development environment with Docker
- ✅ Comprehensive logging and monitoring
- ✅ CI/CD pipeline with GitHub Actions
- ✅ Security scanning and dependency management
- ✅ NEW: RAG retrieval system with semantic search
- ✅ NEW: Intelligent AST-based code chunking
- ✅ NEW: Context assembly with deduplication
- ✅ NEW: MCP tools (search_code, explain_code, find_similar, get_context)
- ✅ NEW: 5x performance improvement in embedding generation
- ✅ Rule-based model routing (Haiku/Sonnet/Opus)
- ✅ Semantic caching for repeated queries
- ✅ Basic cost tracking and reporting
- ✅ Manual escalation mechanism (@haiku/@sonnet/@opus)
- ✅ Cache management and eviction strategies
- ✅ Routing configuration system
- ✅ Security foundations (API keys, rate limiting, auth)
- 📚 Comprehensive setup and usage documentation
- 🎯 Example configurations and tutorials
- ⚡ Performance optimizations
- 🚀 Community launch preparation
- 📊 Simple analytics dashboard
- 🔄 Production migration guides
- Pro tier with advanced features
- Enterprise deployment options
- Team management capabilities
Signal Hub is open source software licensed under the MIT License. See LICENSE for details.
Signal Hub is built on top of excellent open source projects:
- MCP - Model Context Protocol by Anthropic
- ChromaDB - Vector database for embeddings
- Tree-sitter - Code parsing library
Signal Hub - Where your codebase knowledge meets intelligent AI routing.