High-performance semantic search for codebases with 4-stage RAG optimization.
# 1. First-time setup (installs everything)
./setup.sh
# 2. Start services (interactive menu)
./scripts/manage_services.shFirst time? Run ./setup.sh to install dependencies and configure your system.
Daily use? Run ./scripts/manage_services.sh to start/stop services and manage repositories.
- Dependencies: sentence-transformers, redis, postgresql with pgvector
- ONNX Optimization: 3-5x faster embedding generation
- Services: PostgreSQL, Redis, API server
Add to your MCP settings:
{
"mcpServers": {
"repository-memory-bank": {
"command": "python3",
"args": ["/path/to/repository-memory-bank/mcp_server.py"]
}
}
}# Search across repositories
curl -X POST http://localhost:8000/search \
-H "Content-Type: application/json" \
-d '{"query": "authentication functions"}'
# Health check
curl http://localhost:8000/health- Quick Reference - Essential commands
- MCP Tools - All 16 available tools
- API Guide - REST API documentation
- Configuration - Settings and options
- 5x faster search latency (<200ms)
- 80% memory reduction (400MB vs 2GB)
- 4x throughput (20+ concurrent searches)
- 90%+ cache hit rate with intelligent caching
repository-memory-bank/
├── setup.sh # Hardware-optimized installation
├── mcp_server.py # MCP server for Claude
├── src/ # Source code
├── docs/ # All documentation
├── tests/ # All test files
├── scripts/ # Management and utility scripts
│ └── manage_services.sh # Interactive service manager
└── tools/ # Development tools
- Installation issues: Check
docs/guides/TROUBLESHOOTING.md - Service management: Run
./scripts/manage_services.shfor interactive menu - Configuration: Edit
repo_config.yaml
Built with 4-stage RAG optimization: Hybrid Storage → Intelligent Caching → Async Pipeline → ONNX Optimization