yt-dlp-mcp

An MCP server implementation that integrates with yt-dlp, providing video and audio content download capabilities (e.g. YouTube, Facebook, Tiktok, etc.) for LLMs.

Features

Video Metadata: Extract comprehensive video information without downloading content
Subtitles: Download subtitles in SRT format for LLMs to read
Video Download: Save videos to your Downloads folder with resolution control
Audio Download: Save audios to your Downloads folder
Video Search: Search for videos on YouTube using keywords
Privacy-Focused: Direct download without tracking
MCP Integration: Works with Dive and other MCP-compatible LLMs

Installation

Prerequisites

Install yt-dlp based on your operating system:

# Windows
winget install yt-dlp

# macOS
brew install yt-dlp

# Linux
pip install yt-dlp

With Dive Desktop

Click "+ Add MCP Server" in Dive Desktop
Copy and paste this configuration:

{
  "mcpServers": {
    "yt-dlp": {
      "command": "npx",
      "args": [
        "-y",
        "@kevinwatt/yt-dlp-mcp"
      ]
    }
  }
}

Click "Save" to install the MCP server

Tool Documentation

search_videos
- Search for videos on YouTube using keywords
- Inputs:
  - query (string, required): Search keywords or phrase
  - maxResults (number, optional): Maximum number of results to return (1-50, default: 10)
list_subtitle_languages
- List all available subtitle languages and their formats for a video (including auto-generated captions)
- Inputs:
  - url (string, required): URL of the video
download_video_subtitles
- Download video subtitles in any available format. Supports both regular and auto-generated subtitles
- Inputs:
  - url (string, required): URL of the video
  - language (string, optional): Language code (e.g., 'en', 'zh-Hant', 'ja'). Defaults to 'en'
download_video
- Download video to user's Downloads folder
- Inputs:
  - url (string, required): URL of the video
  - resolution (string, optional): Video resolution ('480p', '720p', '1080p', 'best'). Defaults to '720p'
  - startTime (string, optional): Start time for trimming (format: HH:MM:SS[.ms]) - e.g., '00:01:30' or '00:01:30.500'
  - endTime (string, optional): End time for trimming (format: HH:MM:SS[.ms]) - e.g., '00:02:45' or '00:02:45.500'
download_audio
- Download audio in best available quality (usually m4a/mp3 format) to user's Downloads folder
- Inputs:
  - url (string, required): URL of the video
download_transcript
- Download and clean video subtitles to produce a plain text transcript without timestamps or formatting
- Inputs:
  - url (string, required): URL of the video
  - language (string, optional): Language code (e.g., 'en', 'zh-Hant', 'ja'). Defaults to 'en'
get_video_metadata
- Extract comprehensive video metadata without downloading the content
- Returns detailed information including title, description, channel, timestamps, view counts, and more
- Inputs:
  - url (string, required): URL of the video
  - fields (array, optional): Specific metadata fields to extract (e.g., ['id', 'title', 'description', 'channel']). If not provided, returns all available metadata
get_video_metadata_summary
- Get a human-readable summary of key video metadata
- Returns formatted text with title, channel, duration, views, upload date, and description preview
- Inputs:
  - url (string, required): URL of the video

Usage Examples

Ask your LLM to:

"Search for Python tutorial videos"
"Find JavaScript courses and show me the top 5 results"
"Search for machine learning tutorials with 15 results"
"List available subtitles for this video: https://youtube.com/watch?v=..."
"Download a video from facebook: https://facebook.com/..."
"Download Chinese subtitles from this video: https://youtube.com/watch?v=..."
"Download this video in 1080p: https://youtube.com/watch?v=..."
"Download audio from this YouTube video: https://youtube.com/watch?v=..."
"Get a clean transcript of this video: https://youtube.com/watch?v=..."
"Download Spanish transcript from this video: https://youtube.com/watch?v=..."
"Get metadata for this video: https://youtube.com/watch?v=..."
"Show me the title, description, and channel info for this video: https://youtube.com/watch?v=..."
"Get a summary of this video's metadata: https://youtube.com/watch?v=..."
"Extract just the id, title, and view count from this video: https://youtube.com/watch?v=..."

Manual Start

If needed, start the server manually:

npx @kevinwatt/yt-dlp-mcp

Requirements

Node.js 20+
yt-dlp in system PATH
MCP-compatible LLM service

License

MIT

Author

Dewei Yen

YT-DLP下载器