Memograph is a graph-based memory management MCP server designed to solve the LLM 'memory problem' by structuring your personal or organizational knowledge base for smarter semantic search and retrieval. It allows LLMs and AI assistants to access, organize, query, and manage notes or documents in markdown format via knowledge graphs, hybrid retrieval, and semantic search. It exposes a suite of tools (create/read/update/delete/search) for building, linking, and analyzing memories, supporting features like bidirectional linking, bulk creation, analytics, and autonomous memory management. Integrates with popular LLM clients (like Claude Desktop and Cline), multiple model providers (Ollama, OpenAI, Anthropic), and can be used via Python API, CLI, or as an MCP server.
Visit Memograph's official website for product details and getting started.