google-keep-notes-mcp 0.1.2

MCP server for Google Keep

Google Keep MCP

Google Keep MCP is a Model Context Protocol server that lets MCP-capable clients create, read, update, search, archive, and export Google Keep notes.

What this project is

• Installable as a normal Python package
• Usable from Claude Desktop, Cursor, and any other MCP client
• Designed so each user logs in with their own Google Keep account
• Stores login state locally on each user machine

Install

With uv

uv add google-keep-mcp

With pip

pip install google-keep-mcp

From this repo for local testing

cd D:\MyWorld\google-Keep-MCP
uv pip install -e .

Run

After install, run the server command:

google-keep-mcp

For local repo testing you can also run it through uv:

uv run --directory D:\MyWorld\google-Keep-MCP google-keep-mcp

First-time login

Each user must log in once on their own machine.

Run:

google-keep-mcp login

Then enter:

• your Google email
• your Google Keep master token

The login is saved locally in:

• Windows: C:\Users\<you>\.google_keep_mcp\config.json
• macOS/Linux: ~/.google_keep_mcp/config.json

If the token is revoked or removed, the user must log in again.

Claude Desktop config

For local testing, point Claude Desktop to the installed command:

{
  "mcpServers": {
    "google-keep": {
      "command": "D:\\MyWorld\\google-Keep-MCP\\.venv-test\\Scripts\\google-keep-mcp.exe"
    }
  }
}

If the package is installed globally or into the active environment, this also works:

{
  "mcpServers": {
    "google-keep": {
      "command": "google-keep-mcp"
    }
  }
}

Google Tasks reminders

Reminder tools use Google Tasks OAuth credentials.

Place the credentials file here:

C:\Users\<you>\.google_keep_mcp\credentials.json

The reminder token is cached here:

C:\Users\<you>\.google_keep_mcp\token.json

What new users need

• Install the package
• Run google-keep-mcp login
• Paste their own Google Keep master token
• Then use the MCP server from Claude or any other MCP client

Public release note

This package is ready to publish to PyPI, but it is not public until you upload it with your PyPI account.

License

MIT