email-mcp

email-mcp

MCP server that sends outbound emails with automatic single-thread mapping in Gmail, enabling reply-style conversations within a single thread.

Category
访问服务器

README

email-mcp

MCP server that turns the existing email_bridge.py workflow into one outbound email tool with automatic single-thread mapping.

Tool

  • send
    • Sends an email in the mapped Gmail thread.
    • First call creates the thread.
    • Later calls keep replying in that same thread.

Automatic Thread Mapping

Mapping key selection order:

  1. explicit context_id argument
  2. CODEX_THREAD_ID env var
  3. CODEX_SESSION_ID env var
  4. fallback: process-scoped key (proc-<pid>-<random>)

Behavior:

  • One mapping key -> one Gmail thread.
  • Mapping is stored in ~/.codex/email-bridge/mcp-state/thread-map.json.
  • For non-Codex clients, pass context_id in tool calls if you want separate Gmail threads per chat/session.
  • Optional override: set EMAIL_MCP_PROCESS_SESSION_KEY to force a fixed fallback key.

Prerequisites

This server includes a local bridge script at bridge/email_bridge.py (used as fallback when direct SMTP mode is unavailable).

Configure the same mailbox env vars used by that script:

  • CODEX_EMAIL_ADDRESS
  • CODEX_EMAIL_PASSWORD
  • CODEX_EMAIL_TO (or pass to in tool calls)
  • optional SMTP/IMAP host/port vars from the email-bridge docs

Install

cd email-mcp
npm install

Add To Your App (MCP)

This is a local stdio MCP server.

  • command: node
  • args: ["/absolute/path/to/email-mcp/index.js"]
  • env: mailbox credentials + optional tuning flags

Recommended env vars:

  • CODEX_EMAIL_ADDRESS
  • CODEX_EMAIL_PASSWORD
  • CODEX_EMAIL_TO
  • EMAIL_MCP_PREWARM=true
  • EMAIL_MCP_BRIDGE_SCRIPT=/absolute/path/to/email-mcp/bridge/email_bridge.py
  • EMAIL_MCP_PYTHON=python3

Codex (~/.codex/config.toml)

Add this in ~/.codex/config.toml:

[mcp_servers.email]
command = "node"
args = ["/absolute/path/to/email-mcp/index.js"]
tool_timeout_sec = 3600

[mcp_servers.email.env]
CODEX_EMAIL_ADDRESS = "your_email@gmail.com"
CODEX_EMAIL_PASSWORD = "your_app_password"
CODEX_EMAIL_TO = "recipient@example.com"
EMAIL_MCP_PREWARM = "true"
EMAIL_MCP_BRIDGE_SCRIPT = "/absolute/path/to/email-mcp/bridge/email_bridge.py"
EMAIL_MCP_PYTHON = "python3"

tool_timeout_sec can stay high if you want one shared default, but this send-only server does not need long blocking timeouts.

Cursor (~/.cursor/mcp.json or project .cursor/mcp.json)

Add:

{
  "mcpServers": {
    "email-mcp": {
      "command": "node",
      "args": ["/absolute/path/to/email-mcp/index.js"],
      "env": {
        "CODEX_EMAIL_ADDRESS": "your_email@gmail.com",
        "CODEX_EMAIL_PASSWORD": "your_app_password",
        "CODEX_EMAIL_TO": "recipient@example.com",
        "EMAIL_MCP_PREWARM": "true"
      }
    }
  }
}

Then restart Cursor.

Claude Code

Use MCP add with stdio:

claude mcp add --transport stdio \
  --env CODEX_EMAIL_ADDRESS=your_email@gmail.com \
  --env CODEX_EMAIL_PASSWORD=your_app_password \
  --env CODEX_EMAIL_TO=recipient@example.com \
  --env EMAIL_MCP_PREWARM=true \
  email-mcp -- node /absolute/path/to/email-mcp/index.js

If your Claude client uses JSON mcpServers config instead, use the same block shown in the Cursor example.

Notes

  • This MCP server only sends email.
  • Incoming replies should be handled outside MCP by a watcher or intake flow, for example bridge/email_bridge.py watch --once, bridge/email_bridge.py wait, or your higher-level helper scripts.
  • For stable Gmail threading, each mapped session uses one canonical subject base; later subject inputs are ignored for that session.
  • send uses a pooled SMTP transport when SMTP env vars are present, which reduces repeated send overhead.
  • Startup prewarm is enabled by default (EMAIL_MCP_PREWARM=true) and warms the SMTP path in the background.

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