yt

This MCP server, yt-mcp-server, provides YouTube data utilities that you can leverage through the MCP client. It exposes tools for working with YouTube URLs, video IDs, and thumbnails. The available tools include: get_watch_url, which returns the YouTube watch URL for a given video ID and can optionally start at a specified time; get_thumbnail_url, which returns the thumbnail URL for a given video ID; and get_normalized_url, which normalizes various YouTube URL formats to the canonical watch URL and extracts the video ID. These tools can be invoked by the MCP client to transform inputs into consistent YouTube references, making it easier to integrate YouTube data into your workflows or chat-based assistants.

Prerequisites:\n- Python 3.12 or higher\n- UV package manager\n\nInstallation Methods:\n\nMethod 1: Install from GitHub (Recommended)\n\nbash\nuvx --from git+https://github.com/kdr/yt-mcp-server.git server\n\n\nMethod 2: Local Development Setup\n\nbash\ngit clone https://github.com/kdr/yt-mcp-server.git\ncd yt-mcp-server\nuv venv\nsource .venv/bin/activate # On Unix/macOS\n# or\n.venv\Scripts\activate # On Windows\nuv pip install -e .\n\n\nConfigure MCP Client:\n\njson\n{\n "mcpServers": {\n "yt-mcp-server": {\n "command": "uvx",\n "args": [\n "--from",\n "git+https://github.com/kdr/yt-mcp-server.git",\n "server"\n ]\n }\n }\n}\n\n\nFor local development (alternate method):\n\njson\n{\n "mcpServers": {\n "yt-mcp-server": {\n "command": "uv",\n "args": [\n "--directory",\n "/ABSOLUTE/PATH/TO/PARENT/FOLDER/yt-mcp-server/yt_mcp_server",\n "run",\n "main.py"\n ]\n }\n }\n}\n

Notes and tips:\n- The server is Python-based and relies on UV for installation. Ensure you have Python 3.12+ installed.\n- If you run into environment issues, make sure to activate the virtual environment provided by uv venv before installing with pip.\n- The MCP client configuration should reference the git-based install (as shown in the default mcp_config) for smooth updates. For local development, you can point the client to a local directory with the run command as shown in the README.\n- There are three exposed tools: get_watch_url(video_id, start_time?), get_thumbnail_url(video_id), and get_normalized_url(url). Each tool accepts a minimal set of parameters and returns a URL or an extracted video ID as appropriate.\n- If you modify the code, reinstall in editable mode (pip install -e .) to reflect changes without reinstalling.