gannonh/memento-mcp
MCP server that provides Neo4j-backed knowledge graph memory with vector embeddings and semantic retrieval for LLM clients like Claude Desktop and Cursor.

Memento MCP is a Model Context Protocol server that equips LLMs with persistent, scalable long-term memory capabilities. It stores entities and relations in a Neo4j knowledge graph, with vector embeddings enabling semantic search. The system supports temporal awareness, version history, and configurable confidence decay for knowledge aging. It integrates with any MCP-compatible LLM client to provide resilient ontological memory across conversations.
Frequently asked
- What is gannonh/memento-mcp?
- MCP server that provides Neo4j-backed knowledge graph memory with vector embeddings and semantic retrieval for LLM clients like Claude Desktop and Cursor.
- Is memento-mcp open source?
- Yes — gannonh/memento-mcp is open source, released under the MIT license.
- What language is memento-mcp written in?
- gannonh/memento-mcp is primarily written in TypeScript.
- How popular is memento-mcp?
- gannonh/memento-mcp has 425 stars on GitHub.
- Where can I find memento-mcp?
- gannonh/memento-mcp is on GitHub at https://github.com/gannonh/memento-mcp.