knowledge-rag and CoreRag
About knowledge-rag
lyonzin/knowledge-rag
Local RAG System for Claude Code - Hybrid search (Semantic + BM25) with MCP integration
Embeds documents using FastEmbed's ONNX runtime (384-dim BAAI/bge-small-en-v1.5) with cross-encoder reranking and markdown-aware section-based chunking for semantic coherence. Provides 12 MCP tools for document CRUD, URL ingestion, and retrieval evaluation, plus real-time file watching with automatic reindexing and query expansion for security terminology.
About CoreRag
TJ-Neary/CoreRag
Local-first knowledge engine with semantic search, GraphRAG, and Claude Desktop integration via MCP. Features hybrid search (vector + BM25), cross-encoder reranking, multi-format document ingestion (PDF, DOCX, audio, video), and privacy-preserving PII detection. Apple Silicon optimized.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work