
Local persistent semantic memory for AI agents and code copilots.
Visit Copilot Memory MCPCopilot Memory MCP is a local Model Context Protocol (MCP) server that provides persistent, project-scoped semantic memory for AI agents such as GitHub Copilot, Claude, and Cursor. It utilizes an embedded ChromaDB vector database combined with sentence-transformer embeddings to enable semantic search and retrieval-augmented generation (RAG) capabilities, all running locally via Docker with no cloud dependency. Its tools enable agents to create, search, update, delete, and list structured memories, allowing context, decisions, and conventions to persist and be queried across coding sessions, supporting more intelligent, context-aware agent behavior for developers.
Visit Copilot Memory MCP's official website for product details and getting started.