TikTok MCP Service

A Model Context Protocol service for TikTok video discovery and metadata extraction. This service provides a robust interface for searching TikTok videos by hashtags and retrieving trending content, with built-in anti-detection measures and error handling.

Features

• Search videos by hashtags
• Configurable video count per search (default: 30)
• Anti-bot detection measures
• Proxy support
• Automatic API session management
• Rate limiting and error handling
• Health status monitoring

Configuration

The service uses environment variables for configuration. Create a .env file with:

ms_token=your_tiktok_ms_token  # Optional but recommended to avoid bot detection
TIKTOK_PROXY=your_proxy_url    # Optional proxy configuration

Installation and Setup

# Install dependencies
poetry install

# Install browser automation dependencies
poetry run python -m playwright install

# Start the service
poetry run python -m tiktok_mcp_service.main

Claude Desktop Integration

Once your service is running, you can integrate it with Claude Desktop. Since we’re using Poetry for dependency management, make sure to run the MCP CLI commands through Poetry:

# Navigate to the project directory
cd /path/to/tiktok-mcp-service

# Install the service in Claude Desktop with Poetry in editable mode
poetry run mcp install tiktok_mcp_service/main.py --with-editable . -f .env

# Optional: Install with a custom name
poetry run mcp install tiktok_mcp_service/main.py --name "TikTok Video Search" --with-editable . -f .env

After installation, the service will be available in Claude Desktop and will run using Poetry for proper dependency management.

API Endpoints

Health Check

• GET /health - Check service health and API initialization status

  {
    "status": "running",
    "api_initialized": true,
    "service": {
      "name": "TikTok MCP Service",
      "version": "0.1.0",
      "description": "A Model Context Protocol service for searching TikTok videos"
    }
  }
  

Search Videos

• POST /search - Search for videos with hashtags

  {
    "search_terms": ["python", "coding"],
    "count": 30  // Optional, defaults to 30
  }
  

  Response includes video URLs, descriptions, and engagement statistics (views, likes, shares, comments).

Resource Management

• POST /cleanup - Clean up resources and API sessions

Error Handling

The service includes comprehensive error handling for:

• API initialization failures
• Bot detection issues
• Network errors
• Rate limiting
• Invalid search terms

Development

Built with:

• TikTokApi
• FastMCP
• Poetry for dependency management
• Playwright for browser automation

License

MIT