Memory Graph
π§MCP ServerMyangsun/loc-memory-server
|Persistent memory MCP server using a local knowledge graph, allowing Claude to remember user information across chats through entities, relations, and observations with full CRUD and search capabilities.
Overview
Memory Graph is a persistent memory MCP server that uses a local knowledge graph to let Claude remember information about users across chat sessions. It stores knowledge as entities, relations, and observations, providing full CRUD operations and search capabilities so AI assistants can build and maintain a growing understanding of user context over time.
Key Features
- Entity Management β Create, read, and delete entities with unique names, types (person, organization, event), and attached observations
- Relation Tracking β Define directed connections between entities using active-voice relationship types, enabling graph-based knowledge representation
- Observation System β Attach atomic, discrete pieces of information to entities that can be independently added or removed
- Search & Retrieval β Search across the knowledge graph and open specific nodes to recall previously stored information
Who is this for?
Developers who want their AI assistant to maintain persistent memory across conversations β remembering user preferences, project context, team relationships, and other information without needing to be reminded each session. Useful for building personalized AI experiences that improve over time.
Installation
{
"name": "John_Smith",
"entityType": "person",
"observations": ["Speaks f...