Knowledge Graph Memory Server

Example Use Cases

• 1

  Persistently store personalized user facts (identity, preferences, goals) across chat sessions for better context and personalization.

• 2

  Search and retrieve relevant memories by keywords across entity names, types, subdomains, and observation text.

• 3

  Represent and traverse relationships between entities (people, organizations, events) to infer or display connected context.

Description

This MCP server implements a simple knowledge graph-based memory store with entities, relations, and atomic observations, enabling agents to persistently remember user information across conversations. It provides APIs to create/read/delete entities, relations, and observations, plus search and node retrieval endpoints. The server is distributed under the MIT license and can be run via Docker or npx for easy local integration with agents like Claude Desktop.

Capabilities(9 total)