MCP MITM Mem0

Provides an MCP server behind a MITM proxy to store and query Claude conversations via Mem0.

typescript

1

GitHub Stars

typescript

Language

3 months ago

First Indexed

3 weeks ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "terrymunro-mcp-mitm-mem0": {
      "command": "uv",
      "args": [
        "run",
        "mcp-mitm-mem0"
      ],
      "env": {
        "DEBUG": "false",
        "MCP_NAME": "memory_mcp",
        "MITM_HOST": "localhost",
        "MITM_PORT": "8080",
        "MEM0_API_KEY": "YOUR_MEM0_API_KEY",
        "DEFAULT_USER_ID": "default_user"
      }
    }
  }
}

You can run a dedicated MCP server that sits behind a MITM proxy to capture Claude conversations in Mem0 memory and provide memory search, retrieval, and analysis through MCP tools. This setup lets Claude recall past dialogs, track project context, and surface insights from your conversation history.

How to use

Set up and run the memory MCP server, then connect your Claude client through the local MITM proxy to access stored memories and memory-related actions.

How to install

Prerequisites: you need Node.js and the uv tool available in your environment. You also need a Mem0 API key to store memories.

Create the environment configuration and add your Mem0 API key.

cp .env.example .env
# Edit .env and add your MEM0_API_KEY

Additional setup steps

Start the MITM proxy that captures Claude conversations for memory storage.

mitmdump -s memory_addon.py

Start the MCP server that exposes memory tools via MCP.

uv run mcp-mitm-mem0

Create your MCP client configuration to point to the memory MCP server. Use the example below for your .mcp.json.

Configure your MCP client

Add a memory service entry to your MCP configuration so the client can reach the memory MCP server through the dedicated stdio process.

{
  "mcpServers": {
    "memory_mcp": {
      "type": "stdio",
      "command": "uv",
      "args": ["run", "mcp-mitm-mem0"],
      "cwd": "/path/to/mcp-mitm-mem0"
    }
  }
}

MCP Tools

Available tools you can call through the MCP server include searching memories, listing memories, adding memories, deleting memories, analyzing conversations, and suggesting next actions.

MCP Resources

Memory browsing endpoints are available as memory:// URIs to access user memories and recent memories.

Architecture

The Claude client connects through a MITM proxy to the memory service, which stores conversations in Mem0. The MCP Server exposes tools to interact with memories and to analyze conversations, while the Reflection Agent analyzes patterns to provide insights.

Concrete use patterns

Use cases include maintaining personal memory across sessions, preserving project context, and surfacing insights from prior conversations to guide future interactions with Claude.

Development

Run tests and linting as part of your local development workflow.

Configuration

All configuration is driven by environment variables. Set MEM0_API_KEY for Mem0 access, and adjust MITM_HOST, MITM_PORT, and MCP_NAME as needed.

Available tools

search_memories

Search memories using natural language across conversation history

list_memories

List all memories for a given user_id or the default user

add_memory

Manually add memories with optional metadata

delete_memory

Delete a specific memory by memory_id

analyze_conversations

Analyze patterns in recent conversations to surface insights

suggest_next_actions

Suggest next actions based on conversation history