home / mcp / mem0 mcp server

mem0 MCP Server

Forked repo to customize into Project Management purpose.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "kunihiros-mem0-mcp-for-pm": {
      "command": "pipx",
      "args": [
        "run",
        "mem0-mcp-for-pm",
        "--log=off"
      ],
      "env": {
        "MEM0_API_KEY": "{apikey}"
      }
    }
  }
}

mem0 MCP Server for Project Management bridges MCP Host applications with the mem0 cloud service to enable structured project memory management and semantic search for project information. It offers a standardized TOML-based data format, flexible logging, and stdio-based MCP tooling to streamline project memory workflows from your MCP Host.

How to use

You interact with mem0 MCP Server from your MCP Host by running the server through a local MCP client. Use the stdio-based workflow to launch mem0 with the appropriate logging configuration. Store project memories in TOML format so AI systems can extract metadata, enable fast semantic search, and keep your project data organized.

How to install

Prerequisites: ensure you have an environment capable of running Python and pipx. You will execute commands to install and run the MCP server using pipx.

Step-by-step commands:

# Ensure you have pipx installed
python -m pip install --user pipx
python -m pipx ensurepath

# Run mem0 MCP for project management in stdio mode with logging disabled (example)
```
- The final start command is provided in the server configuration examples below. Replace with the appropriate version you intend to run.

Configuration and usage notes

Logging configuration is required when you enable logging. You must specify the log mode and absolute log file path when logging is on. See the examples for exact command-line usage.

Tools are available via the stdio-based MCP protocol and include operations to add, retrieve, search, update, and delete project memories.

Examples and practical setup

"mem0": {
  "command": "pipx",
  "args": ["run", "mem0-mcp-for-pm", "--log=off"],
  "env": {
    "MEM0_API_KEY": "{apikey}"
  }
}
```

```
"mem0": {
  "command": "pipx",
  "args": ["run", "mem0-mcp-for-pm", "--log=on", "--logfile=/workspace/logs/mem0-mcp-server.log"],
  "env": {
    "MEM0_API_KEY": "{apikey}"
  }
}
```
- When logging is enabled, logs are written only to the specified absolute file path. Relative paths or omitting --logfile will cause an error.

Additional configuration details

MCP Host configuration requires explicit log output mode and an absolute log file path when logging is enabled. The following guidelines ensure proper startup and logging behavior.

Security and best practices

Keep your MEM0_API_KEY secret. Use TOML-based templates for project memories to maximize AI extraction and searchability. Maintain consistent field names and include a top-level project and timestamp in every memory entry.

Troubleshooting

If you encounter issues starting the server, verify that the required arguments for logging are present and that the log file path is absolute and writable by the process. Ensure you are using the latest mem0-mcp-for-pm version.

Available tools

add_project_memory

Adds a new project memory entry in TOML format to mem0.

get_all_project_memories

Retrieves all stored project memories.

search_project_memories

Performs a semantic search over project memories.

update_project_memory

Updates an existing project memory entry.

delete_project_memory

Deletes a specific project memory entry.

delete_all_project_memories

Removes all project memories from mem0.