YouTube Transcript MCP Server

What is YouTube Transcript MCP Server?

A high-performance, serverless implementation of a YouTube transcript extraction service using the Model Context Protocol (MCP), deployed on Cloudflare Workers. It enables AI assistants to retrieve transcripts from YouTube videos through a simple API.

How to use YouTube Transcript MCP Server?

Configure Claude Desktop by adding the server’s SSE endpoint to the MCP configuration file using npx mcp-remote, then restart Claude. Invoke the get_transcript tool with a YouTube URL or video ID (and optional language code) to receive the transcript.

Key features of YouTube Transcript MCP Server

• Serverless deployment on Cloudflare’s global edge network
• Supports multiple YouTube URL formats and direct video IDs
• Language selection for transcripts (defaults to English)
• Ultra-fast response times (typically 400–800ms)
• Minimal codebase (less than 300 lines)
• Uses Server-Sent Events (SSE) for streaming connections

Use cases of YouTube Transcript MCP Server

• Get the transcript of any YouTube video for analysis or note-taking
• Summarize key points from a video (e.g., TED talks)
• Integrate transcript extraction into AI-assisted workflows on Claude Desktop

FAQ from YouTube Transcript MCP Server

What are the prerequisites for local development?

Node.js 18 or higher and the Wrangler CLI are required. Install Wrangler globally with npm install -g wrangler.

How do I deploy to Cloudflare?

Run wrangler deploy from the project directory to push the worker to Cloudflare’s edge network.

What transport protocol does the server use?

The server implements Server-Sent Events (SSE) for real‑time streaming connections.

Can I get transcripts in different languages?

Yes, the get_transcript tool accepts an optional lang parameter (default is "en") to specify the language code.

How do I test the server locally?

Start the server with wrangler dev and use the MCP Inspector with SSE transport pointing to http://localhost:8787/sse.