#3333rd of 67 in MCP servers for developers

mcp-memory-service

Self-hosted memory service for AI agents with REST API and MCP

Stars
2k
License
Apache-2.0
Last commit
Oct 2026
Last release
Oct 2026
Language
Python
Stars, last 29 weeks+482
31 Mar 2026: 1,549 stars7 Apr 2026: 1,589 stars14 Apr 2026: 1,639 stars21 Apr 2026: 1,680 stars28 Apr 2026: 1,713 stars5 May 2026: 1,752 stars12 May 2026: 1,804 stars19 May 2026: 1,834 stars26 May 2026: 1,866 stars2 Jun 2026: 1,889 stars9 Jun 2026: 1,889 stars16 Jun 2026: 1,889 stars23 Jun 2026: 1,889 stars30 Jun 2026: 1,889 stars7 Jul 2026: 1,889 stars14 Jul 2026: 1,889 stars21 Jul 2026: 1,889 stars28 Jul 2026: 1,889 stars4 Aug 2026: 1,889 stars11 Aug 2026: 1,889 stars18 Aug 2026: 1,889 stars25 Aug 2026: 1,897 stars1 Sept 2026: 1,916 stars8 Sept 2026: 1,927 stars15 Sept 2026: 1,942 stars22 Sept 2026: 1,952 stars29 Sept 2026: 1,972 stars6 Oct 2026: 1,990 stars13 Oct 2026: 2,014 stars20 Oct 2026: 2,031 stars

1,549 on 31 Mar 20262,031 on 20 Oct 2026

From GitHub's daily star counts.

Checked on GitHub.

Overview

Python service that stores and retrieves agent memories through a REST API, MCP, a CLI and a web dashboard on port 8000. It combines local ONNX embeddings with BM25 search, a typed knowledge graph, and scheduled consolidation, on SQLite by default. Hybrid (Cloudflare or another instance) and Milvus backends are optional.

Who it is for: Developers giving coding assistants or agent pipelines persistent shared memory

Strengths

  • Same store reachable via REST, MCP, CLI and dashboard; no MCP client needed for LangGraph or CrewAI
  • Embeddings run locally with ONNX and SQLite; core operations stay on your machine
  • Backends: SQLite, Cloudflare, Hybrid sync, Milvus; also imports mem0 exports
  • Binds to 127.0.0.1 by default; OAuth 2.0 supported for remote claude.ai and ChatGPT use

Weaknesses

  • Maintained by one person, per the README
  • No Docker or compose file detected in the repository
  • Summaries, bootstrap profiles and quality scoring need extra env flags or an LLM endpoint
  • Bootstrap profiles are not isolated by project

What it needs

  • no GPU
  • Needs SQLite, Cloudflare (optional), Milvus (optional), OpenAI-compatible LLM endpoint (optional)
  • Models: ONNX embeddings (local), OpenAI-compatible endpoints (Ollama, LiteLLM, vLLM) for quality scoring
  • port 8000

Also in MCP servers for developers

See all 67
Also in MCP servers for developers
RankProjectScore
1gitnexusCode knowledge graph exposed to coding agents through MCP tools Live demo ↗ (opens in a new tab)48k starscustom license92 out of 100
2codegraphLocal code knowledge graph that coding agents query over MCP74.6k starsMIT82 out of 100
3n8n-mcpMCP server exposing n8n node docs, templates and workflow management to AI assistants Live demo ↗ (opens in a new tab)23.1k starsMIT78 out of 100
#4serenaMCP server giving coding agents symbol-level code search, editing and refactoring30.6k starscustom license75 out of 100
#5chrome-devtools-mcpMCP server and CLI that lets coding agents control and debug live Chrome54k starsApache-2.073 out of 100