MCP-Mem0: Long-Term Memory for AI Agents

@coleam00

代码仓库

基本信息

分类

记忆与知识

许可证

MIT license

传输方式

stdio

作者

coleam00

关于 MCP-Mem0: Long-Term Memory for AI Agents

MCP server for long term agent memory with Mem0. Also useful as a template to get you started building your own MCP server with Python!

#mcp-mem0#long-term-memory #ai-agents

MCP-Mem0: Long-Term Memory for AI Agents 配置

使用下面的配置,将此服务器添加到你的 MCP 客户端。

{
  "mcpServers": {
    "mcp-mem0": {
      "command": "uv",
      "args": [
        "pip",
        "install",
        "-e",
        "."
      ]
    }
  }
}

MCP-Mem0: Long-Term Memory for AI Agents 工具

工具列表暂未提供

当前收录暂时没有可展示的工具列表，不代表服务没有工具。请查阅项目文档了解支持的能力。

工具是从 README 中自动提取的。维护者可以在 ## Tools 标题下列出工具,即可填充这部分内容。

概览

What is MCP-Mem0?

MCP-Mem0 is an implementation of the Model Context Protocol (MCP) server integrated with Mem0, providing AI agents with persistent memory capabilities. It enables agents to store, retrieve, and search memories using semantic search, serving as a practical template for building custom MCP servers.

How to use MCP-Mem0?

Install dependencies with uv pip install -e . or build a Docker image. Configure a .env file with transport (sse or stdio), LLM provider (OpenAI, OpenRouter, or Ollama), embedding model, and a PostgreSQL database URL. Run the server directly with uv run src/main.py or via Docker. Then integrate with any MCP-compatible client (e.g., Claude Desktop, Windsurf) by adding the server configuration JSON with either SSE or stdio transport.

Key features of MCP-Mem0

Three memory management tools: save, retrieve, and search memories
Semantic search for finding relevant memories
Supports OpenAI, OpenRouter, and Ollama LLM providers
Deployable via Docker (recommended) or with uv
Works with SSE or stdio transport protocols

Use cases of MCP-Mem0

Give AI agents long-term memory that persists across conversations
Retrieve comprehensive context from all stored memories
Find semantically relevant memories for dynamic recall
Build custom MCP servers using this as a reference template

常见问题

What are the prerequisites for running MCP-Mem0?

Python 3.12+, a Supabase or PostgreSQL database for vector storage, API keys for your chosen LLM provider, and optionally Docker.

What memory tools does MCP-Mem0 provide?

It provides three tools: `save_memory` to store information with semantic indexing, `get_all_memories` to retrieve all stored memories, and `search_memories` to find relevant memories via semantic search.

Which transport protocols are supported?

MCP-Mem0 supports both SSE (Server-Sent Events) and stdio transport, configurable via the `TRANSPORT` environment variable.

Where are the memories stored?

Memories are stored as vectors in a PostgreSQL database (Supabase recommended), configured through the `DATABASE_URL` environment variable.

How do I integrate MCP-Mem0 with Windsurf or n8n?

For Windsurf use `serverUrl` instead of `url` in the MCP configuration. For n8n, replace `localhost` with `host.docker.internal` in the SSE URL.

记忆与知识分类下的更多 MCP 服务器

Orcha

westonhancock

Orcha is a unified context layer for AI tools and agents. It stores and indexes and exposes organizational, team, or individual knowledge. Files, structured databases, and connected sources live in one workspace, and age