This MCP server enables Claude Desktop and other MCP-enabled apps to interact with the Douyin video processor. It exposes tools to parse Douyin video information, fetch a no-watermark download link, and extract marketing-friendly or transcript-style text from the video using an AI transcription service. The available tools are: parse_douyin_video_info (parse video metadata), get_douyin_download_link (retrieve no-watermark video URL), and extract_douyin_text (transcribe audio to text). To use it, configure the MCP client to call the extract_douyin_text tool when you want a transcription, or use the info/download tools to retrieve metadata and a download link. The server can handle long audio by chunking when needed, and it supports integration with web UIs via the WebUI component as well as direct MCP tool calls in your workflow.

Prerequisites:

• Python 3.10+ installed on your system
• uv (the Unified Verifier) installed and available in PATH
• FFmpeg installed (for audio/video processing)

Installation steps:

1. Clone the repository git clone https://github.com/yzfly/douyin-mcp-server.git cd douyin-mcp-server

2. Install uv dependencies uv sync

3. Verify installation and basic usage

   Example to list available CLI actions (within the repository context)

   uv run python douyin-video/scripts/douyin_downloader.py --help

4. Configure MCP server (as shown in README) and run via your MCP host using the uvx command configured in mcp_config.

Note: The MCP runtime is typically managed by your MCP host platform. The above steps install and prepare the server so that the uvx-based command can be invoked by the MCP host as configured.

Tips and notes:

• The API key for the AI transcription service is optional in the WebUI configuration but required for extract_douyin_text operations. You can pass API_KEY via environment variable API_KEY or configure it in the MCP config as shown.
• Large audio files are automatically chunked into 9-minute segments for transcription; the results are merged automatically.
• Ensure FFmpeg is installed and available in PATH for audio/video processing during download and transcription workflows.
• The MCP configuration example uses the server name douyin-mcp; adapt the name if you deploy multiple MCP servers.
• If you encounter API rate limits or API key quota issues, consider adjusting chunking behavior or upgrading your API plan with the transcription provider.