WeCom Bot MCP Server

A Model Context Protocol (MCP) compliant server implementation for WeCom (WeChat Work) bot.

English | 中文

Features

• Support for multiple message types:
  • Text messages
  • Markdown messages
  • Image messages (base64)
  • File messages
• @mention support (via user ID or phone number)
• Message history tracking
• Configurable logging system
• Full type annotations
• Pydantic-based data validation

Requirements

• Python 3.10+
• WeCom Bot Webhook URL (obtained from WeCom group settings)

Installation

There are several ways to install WeCom Bot MCP Server:

1. Automated Installation (Recommended)

Using Smithery (For Claude Desktop):

npx -y @smithery/cli install wecom-bot-mcp-server --client claude

Using VSCode with Cline Extension:

1. Install Cline Extension from VSCode marketplace
2. Open Command Palette (Ctrl+Shift+P / Cmd+Shift+P)
3. Search for "Cline: Install Package"
4. Type "wecom-bot-mcp-server" and press Enter

2. Manual Configuration

Add the server to your MCP client configuration file:

// For Claude Desktop on macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
// For Claude Desktop on Windows: %APPDATA%\Claude\claude_desktop_config.json
// For Windsurf: ~/.windsurf/config.json
// For Cline in VSCode: VSCode Settings > Cline > MCP Settings
{
  "mcpServers": {
    "wecom": {
      "command": "uvx",
      "args": [
        "wecom-bot-mcp-server"
      ],
      "env": {
        "WECOM_WEBHOOK_URL": "your-webhook-url"
      }
    }
  }
}

Configuration

Setting Environment Variables

# Windows PowerShell
$env:WECOM_WEBHOOK_URL = "your-webhook-url"

# Optional configurations
$env:MCP_LOG_LEVEL = "DEBUG"  # Log levels: DEBUG, INFO, WARNING, ERROR, CRITICAL
$env:MCP_LOG_FILE = "path/to/custom/log/file.log"  # Custom log file path

Log Management

The logging system uses platformdirs.user_log_dir() for cross-platform log file management:

• Windows: C:\Users\<username>\AppData\Local\hal\wecom-bot-mcp-server\Logs
• Linux: ~/.local/state/hal/wecom-bot-mcp-server/log
• macOS: ~/Library/Logs/hal/wecom-bot-mcp-server

The log file is named mcp_wecom.log and is stored in the above directory.

You can customize the log level and file path using environment variables:

• MCP_LOG_LEVEL: Set to DEBUG, INFO, WARNING, ERROR, or CRITICAL
• MCP_LOG_FILE: Set to a custom log file path

Usage

Once configured, the MCP server runs automatically when your MCP client starts. You can interact with it through natural language in your AI assistant.

Usage Examples

Scenario 1: Send weather information to WeCom

USER: "How's the weather in Shenzhen today? Send it to WeCom"
ASSISTANT: "I'll check Shenzhen's weather and send it to WeCom"
[The assistant will use the send_message tool to send the weather information]

Scenario 2: Send meeting reminder and @mention relevant people

USER: "Send a reminder for the 3 PM project review meeting, remind Zhang San and Li Si to attend"
ASSISTANT: "I'll send the meeting reminder"
[The assistant will use the send_message tool with mentioned_list parameter]

Scenario 3: Send a file

USER: "Send this weekly report to the WeCom group"
ASSISTANT: "I'll send the weekly report"
[The assistant will use the send_file tool]

Scenario 4: Send an image

USER: "Send this chart image to WeCom"
ASSISTANT: "I'll send the image"
[The assistant will use the send_image tool]

Available MCP Tools

The server provides the following tools that your AI assistant can use:

1. send_message - Send text or markdown messages

   • Parameters: content, msg_type (text/markdown), mentioned_list, mentioned_mobile_list

2. send_file - Send files to WeCom

   • Parameters: file_path

3. send_image - Send images to WeCom

   • Parameters: image_path (local path or URL)

For Developers: Direct API Usage

If you want to use this package directly in your Python code (not as an MCP server):

from wecom_bot_mcp_server import send_message, send_wecom_file, send_wecom_image

# Send markdown message
await send_message(
    content="**Hello World!**",
    msg_type="markdown"
)

# Send text message and mention users
await send_message(
    content="Hello @user1 @user2",
    msg_type="text",
    mentioned_list=["user1", "user2"]
)

# Send file
await send_wecom_file("/path/to/file.txt")

# Send image
await send_wecom_image("/path/to/image.png")

Development

Setup Development Environment

1. Clone the repository:

git clone https://github.com/loonghao/wecom-bot-mcp-server.git
cd wecom-bot-mcp-server

2. Create a virtual environment and install dependencies:

# Using uv (recommended)
pip install uv
uv venv
uv pip install -e ".[dev]"

# Or using traditional method
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -e ".[dev]"

Testing

# Run all tests with coverage
uvx nox -s pytest

# Run import tests only
uvx nox -s test_imports

# Run specific test file
uvx nox -s pytest -- tests/test_message.py

# Run tests with verbose output
uvx nox -s pytest -- -v

Code Style

# Check code
uvx nox -s lint

# Automatically fix code style issues
uvx nox -s lint_fix

Building and Publishing

# Build the package
uvx nox -s build

# Publish to PyPI (requires authentication)
uvx nox -s publish

Continuous Integration

The project uses GitHub Actions for CI/CD:

• MR Checks: Runs on all pull requests, tests on Ubuntu, Windows, and macOS with Python 3.10, 3.11, and 3.12
• Code Coverage: Uploads coverage reports to Codecov
• Import Tests: Ensures the package can be imported correctly after installation

All dependencies are automatically tested during CI to catch issues early.

Project Structure

wecom-bot-mcp-server/
├── src/
│   └── wecom_bot_mcp_server/
│       ├── __init__.py
│       ├── server.py
│       ├── message.py
│       ├── file.py
│       ├── image.py
│       ├── utils.py
│       └── errors.py
├── tests/
│   ├── test_server.py
│   ├── test_message.py
│   ├── test_file.py
│   └── test_image.py
├── docs/
├── pyproject.toml
├── noxfile.py
└── README.md

License

This project is licensed under the MIT License - see the LICENSE file for details.

Contact

• Author: longhao
• Email: hal.long@outlook.com