Rag Memory
Provides an advanced MCP server for RAG-enabled memory, leveraging a knowledge graph with vector search capabilities.
About
Rag Memory extends basic memory concepts with semantic search, document processing, and hybrid retrieval to facilitate more intelligent memory management within the Model Context Protocol (MCP) ecosystem. Designed for local operation alongside MCP clients, it uses a SQLite backend with vector operations for fast storage and retrieval of entities, relationships, and observations. It is suitable for intelligent document understanding in applications such as conversational memory or knowledge base construction.
Key Features
- 1 GitHub stars
- Hybrid Search: Combines vector similarity with graph traversal
- Document Processing: RAG-enabled document chunking and embedding
- Vector Search: Semantic similarity search using sentence transformers
- SQLite Backend: Fast local storage with sqlite-vec for vector operations
- Knowledge Graph Memory: Persistent entities, relationships, and observations
Use Cases
- Store, process, and intelligently retrieve research papers
- Remember context across chat sessions with semantic understanding
- Build interconnected knowledge from documents
Sponsored