multiGmailMCP
Connects AI assistants to multiple Gmail accounts simultaneously, enabling search, read, draft, send, and reply operations with per-account permission controls.
README
Multi-Gmail MCP Server
A Model Context Protocol (MCP) server that connects your AI assistant to multiple Gmail accounts simultaneously. Search emails, read content, draft, send, and reply — with per-account permission controls (full, read-only, or draft-only).
1. Setup Google Cloud Project
Since this runs locally, you need your own Google Cloud OAuth credentials:
- Go to the Google Cloud Console.
- Create a new project (e.g., "Multi-Gmail MCP").
- Go to APIs & Services > Library, search for Gmail API, and enable it.
- Go to APIs & Services > OAuth consent screen:
- Choose External user type.
- Fill in the required fields (App name, User support email, Developer contact email).
- Click Save and Continue until you reach the Test users step.
- Add the Gmail addresses you plan to connect as test users (required — Google blocks logins for unlisted users).
- Click Save and Continue.
- Go to APIs & Services > Credentials:
- Click Create Credentials > OAuth client ID.
- Choose Desktop app as the application type.
- Click Create.
- Click Download JSON on your new credential.
- Rename the downloaded file to
credentials.jsonand place it in the project root.
2. Install & Build
npm install
npm run build
3. Authenticate Your Accounts
Authorize each Gmail account you want the server to access:
# Full access (read, compose, send)
npm run auth
# Read-only (no drafting or sending)
npm run auth -- --readonly
# Draft-only (can read and draft, but not send)
npm run auth -- --draftonly
This opens a browser window for Google OAuth. Run the command once per account — credentials are saved locally in tokens.json.
Note: Your Google Cloud project will be in "Testing" status, so Google shows a "Google hasn't verified this app" warning. Click Advanced then Go to [Your App Name] (unsafe) — this is expected since you built the app yourself.
4. Connecting to Claude Desktop
Add the server to your Claude Desktop configuration file.
macOS
Config file location: ~/Library/Application Support/Claude/claude_desktop_config.json
{
"mcpServers": {
"multi_gmail": {
"command": "node",
"args": ["/absolute/path/to/multiGoogleMCP/build/index.js"],
"env": {
"NODE_PATH": "/absolute/path/to/multiGoogleMCP/node_modules"
}
}
}
}
Windows
Config file location: %APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"multi_gmail": {
"command": "node",
"args": ["C:\\absolute\\path\\to\\multiGoogleMCP\\build\\index.js"],
"env": {
"NODE_PATH": "C:\\absolute\\path\\to\\multiGoogleMCP\\node_modules"
}
}
}
}
Replace the paths above with the actual absolute path to your project folder.
After editing the config, restart Claude Desktop for changes to take effect.
5. Connecting to VS Code (Copilot)
Create a .vscode/mcp.json file in your workspace (or add to your user settings):
{
"servers": {
"multi_gmail": {
"command": "node",
"args": ["/absolute/path/to/multiGoogleMCP/build/index.js"],
"env": {
"NODE_PATH": "/absolute/path/to/multiGoogleMCP/node_modules"
}
}
}
}
On Windows, use backslash paths:
{
"servers": {
"multi_gmail": {
"command": "node",
"args": ["C:\\absolute\\path\\to\\multiGoogleMCP\\build\\index.js"],
"env": {
"NODE_PATH": "C:\\absolute\\path\\to\\multiGoogleMCP\\node_modules"
}
}
}
}
VS Code will prompt you to trust the server the first time it starts.
6. Connecting to Google Antigravity
Open the Agent Panel, go to MCP Servers > Manage MCP Servers > Edit configuration to open mcp_config.json, then add:
{
"mcpServers": {
"multi_gmail": {
"command": "node",
"args": ["/absolute/path/to/multiGoogleMCP/build/index.js"],
"env": {
"NODE_PATH": "/absolute/path/to/multiGoogleMCP/node_modules"
}
}
}
}
Use absolute paths. On Windows, use backslash paths (e.g.,
C:\\Users\\...).
7. Connecting to Cursor
Add a new MCP server in Cursor settings:
- Type:
command - Command:
node /absolute/path/to/multiGoogleMCP/build/index.js
If you run into module resolution errors, set the NODE_PATH environment variable to <project-path>/node_modules.
Available Tools
| Tool | Description |
|---|---|
gmail_list_accounts |
List all authenticated accounts |
gmail_search |
Search emails using Gmail search operators |
gmail_read |
Read the full content of an email by ID |
gmail_draft |
Create a draft email |
gmail_send |
Send an email |
gmail_reply |
Reply to an email within its thread (send or draft) |
Content Type Support
gmail_draft, gmail_send, and gmail_reply accept an optional contentType parameter:
| Value | Behavior |
|---|---|
"text" (default) |
Plain text email |
"markdown" |
Body is converted from Markdown to HTML |
"html" |
Body is sent as raw HTML |
Example — drafting a Markdown email:
{
"email": "you@gmail.com",
"to": "recipient@example.com",
"subject": "Weekly Update",
"body": "# Status Report\n\n**Completed:**\n- Feature A\n- Bug fix B\n\n*Next steps:* Feature C",
"contentType": "markdown"
}
The resulting email renders as formatted HTML in Gmail.
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。