ProtonMail MCP Server
An MCP server that connects to ProtonMail via Bridge, enabling AI assistants like Claude to manage your email.
README
ProtonMail MCP Server
An MCP (Model Context Protocol) server that connects to ProtonMail via Bridge, enabling AI assistants like Claude to manage your email.
Features
- Email Management: Read, search, send, reply, forward, and delete emails
- Folder Management: List, create, rename, and delete folders
- Label Management: Apply and remove labels from emails
- Trend Analysis: Analyze email patterns, identify important emails, track sender statistics
Requirements
- Node.js >= 18
- ProtonMail Bridge installed and running
- ProtonMail account (Plus, Unlimited, or Business)
Installation
git clone https://github.com/robotben/protonmail-mcp.git
cd protonmail-mcp
npm install
npm run build
Configuration
-
Copy the example config:
cp config/protonmail.config.example.json config/protonmail.config.json -
Edit
config/protonmail.config.jsonwith your credentials:{ "protonmail": { "auth": { "user": "your-email@protonmail.com", "pass": "your-bridge-password" } } }Note: Use the Bridge password from ProtonMail Bridge app (not your account password).
Running the Server
npm start
The server listens on port 3000 by default. Set the PORT environment variable to change it:
PORT=8080 npm start
MCP endpoint: http://localhost:3000/mcp
Usage with Claude Desktop
The server uses the MCP Streamable HTTP transport (spec 2025-03-26), so configure Claude Desktop with a URL instead of a command:
{
"mcpServers": {
"protonmail": {
"url": "http://localhost:3000/mcp"
}
}
}
Note: Start the server before launching Claude Desktop (
npm start).
Legacy stdio (not supported)
This server no longer supports stdio transport. If you need stdio, use an earlier version.
Available Tools (22)
Email Reading
| Tool | Description |
|---|---|
list_emails |
List emails from a folder with pagination |
get_email |
Get full email content by UID |
get_email_headers |
Get email headers only (lightweight) |
search_emails |
Search emails with query, date range, filters |
get_unread_count |
Get unread count for folders |
mark_as_read |
Mark emails as read |
mark_as_unread |
Mark emails as unread |
Email Sending
| Tool | Description |
|---|---|
send_email |
Send a new email |
reply_to_email |
Reply to an email |
forward_email |
Forward an email |
Folder Management
| Tool | Description |
|---|---|
list_folders |
List all folders |
create_folder |
Create a new folder |
delete_folder |
Delete a folder |
rename_folder |
Rename a folder |
move_emails |
Move emails between folders |
Label Management
| Tool | Description |
|---|---|
list_labels |
List all labels |
create_label |
Create a new label |
apply_labels |
Apply labels to emails |
remove_labels |
Remove labels from emails |
Analytics
| Tool | Description |
|---|---|
analyze_email_trends |
Analyze email patterns over time |
analyze_label_distribution |
Analyze email distribution across folders |
identify_important_emails |
Find important emails based on criteria |
Utility
| Tool | Description |
|---|---|
delete_emails |
Permanently delete emails |
Resources
| URI | Description |
|---|---|
protonmail://inbox/summary |
Inbox statistics |
protonmail://folders |
Folder list with counts |
protonmail://labels |
Label summary |
protonmail://recent |
Recent 24h activity |
protonmail://stats |
Email analytics |
Development
# Build
npm run build
# Watch mode
npm run dev
# Run directly with tsx
npx tsx src/server.ts
License
MIT
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。