Gmail MCP Server
Enables interaction with Gmail through the Gmail API with OAuth2 authentication. Supports reading, sending, searching emails, and managing read status through natural language.
README
Gmail MCP Server
A Model Context Protocol (MCP) server that provides tools for interacting with Gmail through the Gmail API.
Features
- 🔐 OAuth2 Authentication - Secure authentication with Google's OAuth2
- 📧 Read Emails - List and read emails from your Gmail inbox
- ✉️ Send Emails - Send emails through Gmail
- 🔍 Search Emails - Search emails using Gmail's powerful search syntax
- 📌 Mark as Read/Unread - Manage email read status
- 🛡️ Secure - Uses Google's official authentication flow
Quick Start
-
Install dependencies:
pip install -r requirements.txt -
Run the server:
python gmail_mcp_server.py -
First time setup:
- Use the
authenticate_gmailtool to authorize access - This will open your browser for Google OAuth consent
- Use the
Available Tools
authenticate_gmail
Authenticate with Gmail API using OAuth2. Must be run first.
list_emails
List emails from your Gmail inbox.
max_results(optional): Maximum number of emails to return (default: 10)query(optional): Gmail search query
read_email
Read a specific email by its ID.
email_id(required): The ID of the email to read
send_email
Send an email through Gmail.
to(required): Recipient email addresssubject(required): Email subjectbody(required): Email body (plain text)cc(optional): CC email addresses (comma-separated)bcc(optional): BCC email addresses (comma-separated)
search_emails
Search emails using Gmail search syntax.
query(required): Gmail search querymax_results(optional): Maximum number of results (default: 10)
mark_as_read / mark_as_unread
Mark an email as read or unread.
email_id(required): The ID of the email to modify
Gmail Search Syntax
is:unread- Unread emailsfrom:example@gmail.com- Emails from specific sendersubject:important- Emails with "important" in subjecthas:attachment- Emails with attachmentsafter:2024/01/01- Emails after specific date
Configuration
- Credentials file:
gcp-oauth.keys.json(your GCP OAuth credentials - seegcp-oauth.keys.json.example) - Token file:
token.pickle(created automatically after authentication) - MCP config:
mcp_config.json(for MCP client configuration)
Setup Credentials
- Copy
gcp-oauth.keys.json.exampletogcp-oauth.keys.json - Replace the placeholder values with your actual GCP OAuth credentials
- Ensure Gmail API is enabled in your Google Cloud project
Integration
To integrate this server with other projects:
-
Copy the core files:
gmail_mcp_server.py- Main serverrequirements.txt- Dependenciesgcp-oauth.keys.json- Your OAuth credentials
-
Install dependencies:
pip install -r requirements.txt -
Configure MCP client using
mcp_config.json:{ "mcpServers": { "gmail": { "command": "python", "args": ["gmail_mcp_server.py"], "cwd": "/path/to/your/gmail/mcp/server" } } }
Testing
Run the included test to verify functionality:
python direct_test.py
This will test authentication, email listing, searching, and sending.
Security
- OAuth2 credentials are stored securely using Google's official libraries
- Access tokens are refreshed automatically
- All API calls use HTTPS
- Keep
token.picklefile secure
License
This project is provided as-is for educational and integration purposes.
推荐服务器
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。
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。
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。
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。
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。
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。
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。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
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。