mcp-youtube

This MCP server uses yt-dlp to fetch YouTube subtitles and exposes them through the Model Context Protocol so Claude can access video subtitles and context. You can ask Claude to summarize or analyze a video by providing the YouTube URL, for example: 'Summarize the YouTube video https://www.youtube.com/watch?v=abc123'. The server will download available subtitles via yt-dlp and pass the content to Claude as model context for the requested task. Supported capabilities include retrieving subtitles, selecting language, and generating a summary or analysis based on the video transcript. Ensure yt-dlp is installed and available in your environment so the server can fetch subtitles on demand.

Prerequisites:

• Node.js and npm installed on your system
• yt-dlp installed and accessible in your PATH (e.g., via Homebrew on macOS or WinGet on Windows)

Installation steps:

1. Install yt-dlp (if not already installed):

   • macOS: brew install yt-dlp
   • Windows: winget install yt-dlp
   • Linux: typically available via your distro's package manager or via python -m pip install -U yt-dlp

2. Install the MCP server via mcp-installer (or your preferred MCP workflow):

   • Install mcp-installer if you don't have it: npm i -g mcp-installer
   • Install this MCP server package: mcp-installer install @anaisbetts/mcp-youtube

3. Run the MCP server:

   • Run using npx (as intended by the package): npx -y @anaisbetts/mcp-youtube

Note: Ensure your environment variables are set if the server requires the path to yt-dlp or other tools. You can configure YTDLP_PATH or other relevant vars as needed by your setup.

Tips:

• The server relies on yt-dlp to fetch YouTube subtitles; ensure it is up to date for best subtitle support.
• If you encounter issues with subtitle languages, check that the desired language code is supported by yt-dlp and request subtitles accordingly.
• When debugging, run the server in a verbose or debug mode if available and verify that npx can resolve @anaisbetts/mcp-youtube from your environment.
• If using behind a firewall or proxy, ensure yt-dlp and Node can access YouTube endpoints.
• Environment variables: YTDLP_PATH can be set to the explicit path to the yt-dlp executable if it isn’t in your system PATH.