Google Tasks MCP Server

AI assistants and Google Tasks integration through Model Context Protocol.

AIアシスタントとGoogle Tasksを連携させるModel Context Protocol実装。

目次 / Table of Contents

• English
• 日本語

English

Google Tasks MCP Server

An implementation for interacting with Google Tasks API through Model Context Protocol (MCP). This server allows you to use Google Tasks task management features from MCP-compatible clients such as Cursor.

This project was created to practice implementing an MCP server, with the goal of enabling AI assistants to interact with Google Tasks management features through the MCP protocol.

Potential Issues

This project was created for learning purposes, and there may be issues that I'm not aware of due to my limited knowledge.

Please use at your own risk. The developer cannot take responsibility for any problems that may occur.

If you discover any issues, please create an Issue or suggest a fix via Pull Request. See the "Contributing" section below for details.

Contributing

Contributions to this project are welcome. Bug reports, feature suggestions, pull requests, or any other form of participation are greatly appreciated.

For details, please refer to CONTRIBUTING.md.

Features

This MCP server provides the following functions:

• Fetch task list collections
• Get tasks within a specific task list
• Create new tasks
• Mark tasks as completed
• Delete tasks

Setup Guide

1. Prerequisites

• Node.js 16 or higher
• npm (included with Node.js)
• Google Cloud Project account

2. Project Preparation

# Clone the repository (or obtain the project by any method)
git clone <repository URL>
cd mcp-google-tasks

# Install dependencies
npm install

3. Google Cloud Project Configuration

1. Access the Google Cloud Console
2. Create a new project (or select an existing project)
3. From the left menu, select "APIs & Services" → "Library"
4. Search for "Tasks API" in the search box
5. Select "Google Tasks API" and click "Enable"

4. Create OAuth 2.0 Credentials

1. From the left menu, select "APIs & Services" → "Credentials"
2. Click "Create Credentials" → "OAuth client ID"
3. Select "Desktop app" as the application type
4. Enter a name (e.g., "Google Tasks MCP Client") and click "Create"
5. Download the displayed credentials (JSON file)

5. Place Credentials

1. Rename the downloaded JSON file to credentials.json
2. Place this file in the project's root directory (same level as package.json)

mcp-google-tasks/
├── src/
├── package.json
├── credentials.json  <- place here
└── ...

6. Build and Run

# Build TypeScript
npm run build

# Run the server
node build/index.js

On first run, a browser window will open asking for Google account authentication. After authentication is complete, a token.json file will be generated, and this token will be used automatically for subsequent runs.

Integration with Cursor

To use this server with MCP-compatible IDEs like Cursor, configure as follows:

For Cursor

1. Navigate to Cursor's configuration directory (typically ~/.cursor)
2. Create or edit the mcp.json file

{
  "mcpServers": {
    "gtasks": {
      "command": "node",
      "args": [
        "/full/path/mcp-google-tasks/build/index.js"
      ]
    }
  }
}

Replace /full/path/ with the actual full path to your project. Example: /home/username/projects/mcp-google-tasks/build/index.js

Available Tools

list_tasklists

Lists all task lists.

Parameters: none

Example:

list_tasklists()
=> Task list collection (JSON format)

list_tasks

Lists tasks within a specific task list.

Parameters:

• taskListId: Task list ID (obtainable from list_tasklists)

Example:

list_tasks(taskListId: "YOUR_TASK_LIST_ID")
=> Tasks in the specified task list (JSON format)

create_task

Cre

... View full README on GitHub