mcp-memory-service

mcp-memory-service

Provides persistent memory services for AI agents, supporting REST API, knowledge graphs, and autonomous integration.

MCPDevOpen source
Type
MCP
Transport
http
Open source
Yes
GitHub Stars
★ 1.9k
Source
mcp-github

Overview

mcp-memory-service is an open-source persistent memory backend designed specifically for AI agents. It offers multiple transport methods including REST API, MCP, OAuth, and CLI, and supports knowledge graphs and autonomous integration. AI agents can store decisions, share causal chains, and retrieve context in under 5ms without relying on cloud services or paying API fees. Compatible with various frameworks such as LangGraph, CrewAI, and AutoGen. With simple installation and configuration, it can be quickly integrated and used.

Capabilities

  • REST API support for 76 endpoints
  • Shared causal chains via knowledge graph
  • Automatic memory tagging with X-Agent-ID header
  • conversation_id bypasses duplicate elimination
  • SSE events for real-time notifications
  • Local running of embedded models

Use cases

Shared memory across multi-agent clustersPersistent memory across sessionsAutonomous integration of old memoriesDirect use in browser with Claude

Setup

Requires: Python 环境API Key
pip install mcp-memory-service; MCP_ALLOW_ANONYMOUS_ACCESS=true memory server --http

This information was compiled by AI from public sources and may contain inaccuracies — please refer to the source.

FAQ

How to use mcp-memory-service in Claude?

Set up Remote MCP and expose the service via Cloudflare Tunnel, then add a connector in Claude.

Does it support other frameworks?

Yes, it supports multiple frameworks including LangGraph, CrewAI, AutoGen, and more.

Related skills