mcp-gmail

mcp-gmail

An MCP server that provides Gmail access for LLMs, enabling email composition, sending, searching, and label management via the Gmail API.

Category
访问服务器

README

MCP Gmail Server

A Model Context Protocol (MCP) server that provides Gmail access for LLMs, powered by the MCP Python SDK.

Features

  • Expose Gmail messages as MCP resources
  • Provide tools for composing, sending, and managing emails
  • OAuth 2.0 authentication with Google's Gmail API

Prerequisites

  • Python 3.10+
  • Gmail account with API access
  • uv for Python package management (recommended)

Setup

1. Install dependencies

Install project dependencies (uv automatically creates and manages a virtual environment)

uv sync

2. Configure Gmail OAuth credentials

There's unfortunately a lot of steps required to use the Gmail API. I've attempted to capture all of the required steps (as of March 28, 2025) but things may change.

Google Cloud Setup

  1. Create a Google Cloud Project

    • Go to Google Cloud Console
    • Click on the project dropdown at the top of the page
    • Click "New Project"
    • Enter a project name (e.g., "MCP Gmail Integration")
    • Click "Create"
    • Wait for the project to be created and select it from the dropdown
  2. Enable the Gmail API

    • In your Google Cloud project, go to the navigation menu (≡)
    • Select "APIs & Services" > "Library"
    • Search for "Gmail API"
    • Click on the Gmail API card
    • Click "Enable"
  3. Configure OAuth Consent Screen

    • Go to "APIs & Services" > "OAuth consent screen"
    • You will likely see something like "Google Auth Platform not configured yet"
      • Click on "Get started"
    • Fill in the required application information:
      • App name: "MCP Gmail Integration"
      • User support email: Your email address
    • Fill in the required audience information:
      • Choose "External" user type (unless you have a Google Workspace organization)
    • Fill in the required contact information:
      • Your email address
    • Click "Save and Continue"
    • Click "Create"
  4. Create OAuth Credentials

    • Go to "APIs & Services" > "Credentials"
    • Click "Create Credentials" > "OAuth client ID"
    • Choose "Desktop app" as the application type
    • Enter a name (e.g., "MCP Gmail Desktop Client")
    • Click "Create"
    • Click "Download JSON" for the credentials you just created
    • Save the file as credentials.json in your project root directory
  5. Add scopes

    • Go to "APIs & Services" > "OAuth consent screen"
    • Go to the "Data Access" tab
    • Click "Add or remove scopes"
    • Search for the Gmail API
    • Select the scope for .../auth/gmail.modify which grants permission to "Read, compose, and send emails from your Gmail account"
    • Click update
    • Click save

Verify that you've set up your OAuth configuration correctly by running a simple test script.

uv run python scripts/test_gmail_setup.py

You should be able to see usage metrics at https://console.cloud.google.com/apis/api/gmail.googleapis.com/metrics

3. Run the server

Development mode:

uv run mcp dev mcp_gmail/server.py

This will spin up an MCP Inspector application that you can use to interact with the MCP server.

Or install for use with Claude Desktop:

uv run mcp install \
    --with-editable .
    --name gmail \
    --env-var MCP_GMAIL_CREDENTIALS_PATH=$(pwd)/credentials.json \
    --env-var MCP_GMAIL_TOKEN_PATH=$(pwd)/token.json \
    mcp_gmail/server.py

[!NOTE] If you encounter an error like Error: spawn uv ENOENT when spinning up Claude Desktop and initializing the MCP server, you may need to update your claude_desktop_config.json to provide the absolute path to uv. Go to Claude Desktop -> Settings -> Developer -> Edit Config.

{
  "mcpServers": {
    "gmail": {
      "command": "~/.local/bin/uv",
    }
  }
}

Development

Linting and Testing

Run linting and formatting:

# Format code
uv run ruff format .

# Lint code with auto-fixes where possible
uv run ruff check --fix .

# Run tests
uv run pytest tests/

Pre-commit Hooks

This project uses pre-commit hooks to ensure code quality. The hooks automatically run before each commit to verify code formatting and linting standards.

Install the pre-commit hooks:

pre-commit install

Run pre-commit manually on all files:

pre-commit run --all-files

Usage

Once running, you can connect to the MCP server using any MCP client or via Claude Desktop.

Available Resources

  • gmail://messages/{message_id} - Access email messages
  • gmail://threads/{thread_id} - Access email threads

Available Tools

  • compose_email - Create a new email draft
  • send_email - Send an email
  • search_emails - Search for emails with specific filters (from, to, subject, dates, etc.)
  • query_emails - Search for emails using raw Gmail query syntax
  • get_emails - Retrieve multiple email messages by their IDs
  • list_available_labels - Get all available Gmail labels
  • mark_message_read - Mark a message as read
  • add_label_to_message - Add a label to a message
  • remove_label_from_message - Remove a label from a message

Environment Variables

You can configure the server using environment variables:

  • MCP_GMAIL_CREDENTIALS_PATH: Path to the OAuth credentials JSON file (default: "credentials.json")
  • MCP_GMAIL_TOKEN_PATH: Path to store the OAuth token (default: "token.json")
  • MCP_GMAIL_MAX_RESULTS: Default maximum results for search queries (default: 10)

License

MIT

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选