MCP server youtube transcript agent

YouTube Transcript Server

A Model Context Protocol server that enables retrieval of transcripts from YouTube videos. This server provides direct access to video captions and subtitles through a simple interface.

Installing via Smithery

To install YouTube Transcript Server for Claude Desktop automatically via Smithery:

npx -y @smithery/cli install @kimtaeyoon83/mcp-server-youtube-transcript --client claude

Components

Tools

• get_transcript

  • Extract transcripts from YouTube videos
  • Inputs:
    • url (string, required): YouTube video URL, Shorts URL, or video ID
    • lang (string, optional, default: "en"): Language code for transcript (e.g., 'ko', 'en'). Automatically falls back to available languages if requested language is not found.
    • include_timestamps (boolean, optional, default: false): Include timestamps in output (e.g., '[0:05] text')
    • strip_ads (boolean, optional, default: true): Filter out sponsorships, ads, and promotional content from transcript based on chapter markers

• analyze_video (optional, requires a TwelveLabs API key)

  • Analyze a video with TwelveLabs Pegasus, a video-understanding model. Unlike get_transcript, this reasons over what is shown on screen, so it produces useful summaries and answers even for videos with little or no speech (demos, gameplay, b-roll, music videos).
  • Inputs:
    • url (string, required): A publicly reachable direct video URL (e.g. an .mp4/.mov/.webm link or a pre-signed URL). TwelveLabs fetches the file server-side, so a YouTube watch page URL will not work — it serves HTML, not a raw video stream.
    • prompt (string, optional): Instruction or question for the model (e.g. "Summarize this video in 3 sentences" or "What products are shown?"). Defaults to a general summary.
    • model (string, optional, default: "pegasus1.2"): Pegasus model to use ("pegasus1.2" or "pegasus1.5").
    • max_tokens (number, optional, default: 2048): Maximum response length in tokens.
  • Requires the TWELVELABS_API_KEY environment variable. The transcript tool works without it; analysis is fully opt-in. Grab a free key at twelvelabs.io — there's a generous free tier.

Key Features

• Support for multiple video URL formats (including YouTube Shorts)
• Language-specific transcript retrieval with automatic fallback
• Optional timestamps for referencing specific moments
• Built-in ad/sponsorship filtering (enabled by default)
• Zero external dependencies for transcript fetching
• Detailed metadata in responses

Configuration

To use with Claude Desktop, add this server configuration:

{
  "mcpServers": {
    "youtube-transcript": {
      "command": "npx",
      "args": ["-y", "@kimtaeyoon83/mcp-server-youtube-transcript"]
    }
  }
}

To enable the optional analyze_video tool, add your TwelveLabs API key:

{
  "mcpServers": {
    "youtube-transcript": {
      "command": "npx",
      "args": ["-y", "@kimtaeyoon83/mcp-server-youtube-transcript"],
      "env": {
        "TWELVELABS_API_KEY": "your-twelvelabs-api-key"
      }
    }
  }
}

Install via tool

mcp-get A command-line tool for installing and managing Model Context Protocol (MCP) servers.

npx @michaellatman/mcp-get@latest install @kimtaeyoon83/mcp-server-youtube-transcript

Awesome-mcp-servers

awesome-mcp-servers A curated list of awesome Model Context Protocol (MCP) servers.

Development

Prerequisites

• Node.js 18 or higher
• npm or yarn

Setup

Install dependencies:

npm install

Build the server:

npm run build

For development with auto-rebuild:

npm run watch

Testing

npm test

Debugging

Since MCP servers communicate over stdio, debugging can be challenging. We recommend using the MCP Inspector for development:

npm run inspector

Running evals

The evals package loads an mcp client that then runs the index.ts file, so there is no need to rebuild between tests. You can load environment variables by prefixing the npx command. Full documentation can be found here.

OPENAI_API_KEY=your-key  npx mcp-eval src/evals/evals.ts src/index.ts

Error Handling

The server implements robust error handling for common scenarios:

• Invalid video URLs or IDs
• Unavailable transcripts
• Language availability issues
• Network errors

Usage Examples

1. Get transcript by video URL:

await server.callTool("get_transcript", {
  url: "https://www.youtube.com/watch?v=VIDEO_ID",
  lang: "en"
});

2. Get transcript by video ID:

await server.callTool("get_transcript", {
  url: "VIDEO_ID",
  lang: "ko"
});

3. Get transcript from YouTube Shorts:

await server.callTool("get_transcript", {
  url: "https://www.youtube.com/shorts/VIDEO_ID"
});

4. Get transcript with timestamps:

await server.callTool("get_transcript", {
  url: "VIDEO_ID",
  include_timestamps: true
});

5. Get raw transcript without ad filtering:

await server.callTool("get_transcript", {
  url: "VIDEO_ID",
  strip_ads: false
});

6. How to Extract YouTube Subtitles in Claude Desktop App

chat: https://youtu.be/ODaHJzOyVCQ?si=aXkJgso96Deri0aB Extract subtitles

Security Considerations

The server:

• Validates all input parameters
• Handles YouTube API errors gracefully
• Implements timeouts for transcript retrieval
• Provides detailed error messages for troubleshooting

License

This MCP server is licensed under the MIT License. See the LICENSE file for details.