bzurkowski/mcp-youtube

✨ YouTube MCP Server ✨

A Model Context Protocol (MCP) server providing structured access to YouTube data for AI assistants and applications. It lets AI systems interact with YouTube content through a standardized interface. The server offers tools for searching videos, retrieving channel information, accessing transcripts, and analyzing comments.

Table of Contents

• Features
• Installation
  • Prerequisites
  • Running with Docker
• Configuration
  • Cursor AI
• Tools
  • search_videos
  • list_channels
  • list_playlist_videos
  • get_video_details
  • list_video_comments
  • get_video_transcript
  • list_video_transcripts
  • translate_video_transcript

Features

• Video Search - Search for videos across YouTube by keywords and filters
• Channel Information - Retrieve details about YouTube channels by ID, handle, or username
• Playlist Access - Get videos from public YouTube playlists
• Video Metadata - Access detailed metadata for any YouTube video
• Comment Retrieval - Fetch comments from YouTube videos
• Video Transcripts - Get and list video transcripts in multiple languages
• Transcript Translation - Translate video transcripts to different languages

Installation

Prerequisites

• Docker
• Google API credentials for YouTube Data API

Running with Docker

Build the Docker image:

docker build -t mcp-youtube .

Run the container:

docker run -p 8000:8000 -e YOUTUBE_API_KEY=<your_api_key> mcp-youtube

Configuration

Cursor AI

To use this MCP server with Cursor AI, add the following configuration to your MCP settings:

{
  "mcp": {
    "tools": [
      {
        "name": "youtube",
        "server": {
          "url": "http://localhost:8000"
        }
      }
    ]
  }
}

Tools

The following tools are available for interacting with YouTube:

search_videos

Search for videos on YouTube.

Arguments:

• query (required): The search query
• max_results (optional, default: 10): Maximum number of results to return
• order (optional, default: "relevance"): Order of the results ('relevance', 'date', 'rating', 'viewCount')
• region_code (optional): ISO 3166-1 alpha-2 country code (e.g., 'US', 'GB')
• language (optional): ISO 639-1 language code (e.g., 'en', 'fr')
• published_after (optional): Only return videos published after this datetime (RFC 3339 format)
• published_before (optional): Only return videos published before this datetime (RFC 3339 format)

list_channels

List YouTube channel details.

Arguments:

• One of the following is required:
  • channel_id: The YouTube channel ID to retrieve details for
  • handle: The YouTube handle to retrieve details for
  • username: The YouTube username to retrieve details for
• max_results (optional, default: 10): Maximum number of channels to return

list_playlist_videos

List videos from a specific YouTube playlist.

Arguments:

• playlist (required): The playlist ID or URL to list videos from
• max_results (optional, default: 10): Maximum number of videos to return

get_video_details

Get details for a specific YouTube video.

Arguments:

• video (required): The video ID or URL to get details for

list_video_comments

Get comments for a specific YouTube video.

Arguments:

• video (required): The video ID or URL to get comments for
• max_results (optional, default: 20): Maximum number of comments to return
• order (optional, default: "relevance"): Order of the comments ('relevance' or 'time')

get_video_transcript

Get transcript for a specific YouTube video.

Arguments:

• video (required): The video ID or URL to get transcript for
• language_code (required): ISO 639-1 language code of the transcript (e.g., 'en', 'fr')
• preserve_formatting (optional, default: false): Whether to preserve HTML formatting in the transcript
• include_timestamps (optional, default: false): Whether to include timestamps in the transcript

list_video_transcripts

List available transcripts for a specific YouTube video.

Arguments:

• video (required): The video ID or URL to list transcripts for

translate_video_transcript

Translate transcript for a specific YouTube video to the specified language.

Arguments:

• video (required): The video ID or URL to translate transcript for
• language_code (required): ISO 639-1 language code of the translation (e.g., 'en', 'fr').
• preserve_formatting (optional, default: false): Whether to preserve HTML formatting in the transcript