Allows AI agents to interact with WeChat, including sending text messages, files, managing contacts and groups, and retrieving messages.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@wechat-mcp send a message to Alice saying I'm on my way"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
让 Codex 接管微信 — 通过 MCP 协议将微信能力暴露为 AI 可调用的工具。
┌─────────────────────────────────────────────┐
│ Codex │
│ ┌──────────────────────────────┐ │
│ │ WeChat MCP Server │ │
│ │ (暴露 10+ 个工具给 Codex) │ │
│ └───────┬──────────────────────┘ │
│ │ stdio / JSON-RPC │
│ ┌───────▼──────────────────────┐ │
│ │ Wechaty + Puppet │ │
│ │ (微信协议桥接层) │ │
│ └───────┬──────────────────────┘ │
│ │ 注入/协议 │
│ ┌───────▼──────────────────────┐ │
│ │ 微信 (桌面版 / iPad 协议) │ │
│ └──────────────────────────────┘ │
└─────────────────────────────────────────────┘Related MCP server: weixin-mcp-server
# 1. 安装依赖
bash scripts/setup.sh
# 2. 启动（扫码登录微信）
npm start
# 3. 注册到 Codex
codex plugin add wechat-mcp@personal
# 4. 在 Codex 中即可使用 WeChat 工具工具 | 描述 |
| 获取登录状态和 Bot 信息 |
| 发送文本消息 |
| 发送文件/图片 |
| 获取联系人列表（支持搜索） |
| 获取群聊列表 |
| 获取最近消息 |
| 搜索历史消息 |
| 获取联系人详情 |
| 获取自身信息 |
| 搭建指南和故障排查 |
协议: MCP (Model Context Protocol) over stdio
微信桥接: Wechaty + wechaty-puppet-wechat (UOS)
运行时: Node.js 18+
通信: JSON-RPC 2.0
wechat-mcp/
├── .codex-plugin/
│ └── plugin.json # 插件清单
├── .mcp.json # MCP Server 配置
├── mcp/
│ ├── index.mjs # MCP Server 入口
│ └── wechaty-bridge.mjs # Wechaty 桥接封装
├── scripts/
│ └── setup.sh # 一键安装/检查脚本
├── skills/
│ └── SKILL.md # Codex 技能定义
├── package.json # Node.js 依赖
└── README.mdMIT
获取指定联系人的详细信息，包括昵称、备注名、地区、个性签名等。
Inputs: contactId (required).
获取所有联系人或按关键词搜索联系人。返回昵称、备注、ID 等信息。
Inputs: query.
获取最近收到的消息列表。默认返回最近 50 条。
Inputs: limit.
获取所有群聊列表，包含群名称和成员数。
Inputs: no named parameters.
按关键词搜索历史消息记录，支持按文本内容和发送者搜索。
Inputs: keyword (required), limit.
获取当前登录微信账号的个人信息。
Inputs: no named parameters.
向联系人或群聊发送文件或图片，需要提供本地文件的绝对路径。
Inputs: contact (required), filePath (required).
向联系人或群聊发送文本消息。contact 参数可以是联系人的 ID、昵称或群名称。
Inputs: contact (required), message (required).
获取 Wechaty 环境搭建指南、依赖状态和故障排查建议。
Inputs: no named parameters.
获取微信登录状态、Bot 运行信息和 Wechaty 初始化状态
Inputs: no named parameters.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 10 tools
Each tool targets a distinct action or resource: sending text/file, retrieving contacts/rooms/messages, searching history, and fetching status/self information. Even the seemingly similar contacts list and contact info tools are clearly differentiated by scope (list/search vs detailed profile).
All tools share the 'wechat_' prefix, and most follow a verb_noun pattern (wechat_send_text, wechat_get_contacts). However, several tools like wechat_status, wechat_contact_info, wechat_self_info, and wechat_setup_guide lack a leading verb, creating a mixed but still readable convention.
With 10 tools, the server is well-scoped for a WeChat bot, covering messaging, contacts, rooms, message history, and account info without unnecessary expansion. This falls comfortably within the ideal range for a domain-specific MCP server.
The tool set covers the core WeChat operations: sending text and files, listing contacts and rooms, retrieving recent messages, searching history, and accessing self/contact details. Minor gaps exist, such as no explicit tool for group member management or real-time event subscription, but these are not essential for typical bot workflows.
Email infrastructure for AI agents — send, receive, search, and reply to email over MCP.
Connect any AI agent to 1,000+ apps and 27,000+ actions through one remote MCP server (OAuth).
Drive WhatsApp from any MCP client: pair devices, send text and media, manage contacts and groups.
MCP server connecting AI agents to 100+ apps (Gmail, Slack, Notion, GitHub) via one-click OAuth.
