YouTube Knowledge MCP

A Model Context Protocol (MCP) server that gives AI assistants the ability to search, analyze, and extract knowledge from YouTube videos. Works with Claude Desktop, Claude Code, Claude.ai, Cursor and any MCP-compatible client.

Supports both local (stdio) and remote (Streamable HTTP) transports.

Features

• Search videos by keyword or phrase

• Fetch videos from playlists or channels

• Get video info (title, channel, duration, views, likes, comments, description, tags)

• Extract transcripts (auto-generated or manual captions)

• Get chapters (timestamps and structure)

• Get comments (top comments sorted by popularity)

• Get channel info (name, handle, subscriber count, description)

• List formats (available download resolutions and codecs)

• Download videos with quality selection (local mode only)

• Save to library (summaries, notes, skills) (local mode only)

• List library with tag filtering (local mode only)

Prerequisites

• Node.js 20+

• yt-dlp: brew install yt-dlp (macOS) or see installation guide

Installation

Via npm (Recommended)

npm install -g youtube-knowledge-mcp

Via npx (no installation)

Configure directly with npx (see Configuration section).

From source

git clone https://github.com/teobouancheau/youtube-knowledge-mcp.git
cd youtube-knowledge-mcp
npm install
npm run build

Configuration

Local (stdio) -- Claude Desktop, Claude Code, Cursor

Quick Start with npx

{
  "mcpServers": {
    "youtube-knowledge": {
      "command": "npx",
      "args": ["-y", "youtube-knowledge-mcp"]
    }
  }
}

With Global Installation

npm install -g youtube-knowledge-mcp

{
  "mcpServers": {
    "youtube-knowledge": {
      "command": "youtube-knowledge-mcp"
    }
  }
}

Configuration File Locations

Restart your client after updating configuration.

Remote (HTTP) -- Claude.ai, Claude Mobile, Custom Connectors

The server supports Streamable HTTP transport for remote access via Claude's official connectors.

Self-hosted

npm run build
npm run start:http

The server listens on PORT (default 3000). Set PORT environment variable to change.

Docker

npm run build
docker build -t youtube-knowledge-mcp .
docker run -p 3000:10000 youtube-knowledge-mcp

Deploy to Render

1. Push to GitHub

2. Create a Web Service on Render with Docker runtime

3. Render sets PORT automatically

4. Add the Render URL as a custom connector in Claude.ai > Settings > Connectors

Connect via Claude.ai

1. Go to Settings > Connectors

2. Click Add custom connector

3. Enter your server URL (e.g., https://your-app.onrender.com/mcp)

4. Click Add

MCP Tools

Remote + Local (10 tools)

search_videos

Search YouTube for videos by keyword or phrase.

Returns: video IDs, titles, durations, channels, view counts, URLs

fetch_videos

List videos from a YouTube playlist or channel.

get_video_info

Get detailed metadata for a single YouTube video.

Returns: title, channel, duration, upload date, view count, like count, comment count, description, tags, thumbnail URL

get_transcript

Extract the full transcript from a YouTube video.

Returns: plain text transcript with word count and detected language

get_chapters

Extract chapter markers and timestamps from a YouTube video.

Returns: chapter titles with start and end times. Empty list if no chapters found.

get_comments

Get top comments from a YouTube video sorted by popularity.

Returns: author, text, like count, pinned status

get_channel_info

Get metadata for a YouTube channel.

Returns: channel name, handle, subscriber count, description, channel URL

search_channels

Search YouTube for channels by keyword or phrase.

Returns: channel names, handles, subscriber counts, descriptions, URLs

get_playlist_info

Get metadata for a YouTube playlist.

Returns: title, channel, video count, last updated date, description

list_formats

List all available download formats for a YouTube video.

Returns: format IDs, extensions, resolutions, FPS, codecs, file sizes

Local Only (3 tools, stdio mode)

These tools operate on the local filesystem and are only available in stdio mode.

download_video

Download a YouTube video to local disk.

save_to_library

Save a summary or skill note to your local YouTube knowledge library.

list_library

List all saved items in your local knowledge library.

Library Storage

Content is stored in ~/.youtube-knowledge/:

~/.youtube-knowledge/
├── transcripts/          # Cached transcripts
│   └── {video_id}.txt
├── library/              # Saved content
│   └── {video_id}/
│       ├── metadata.json
│       ├── summary.md
│       └── skill.md
├── downloads/            # Downloaded videos
│   └── {video_title}.{ext}
└── index.json            # Searchable index

Usage Examples

Search and analyze

"Search YouTube for 'transformer architecture explained' and summarize the top result"

Explore a channel

"Show me the latest videos from @ThePrimeagen and get the chapters for the most recent one"

Deep content analysis

"Get the transcript and chapters for this video, then create a structured summary"

Audience sentiment

"What are the top comments on this video? What does the audience think?"

Download a video (local only)

"Download this video in 1080p: https://youtube.com/watch?v=ABC123"

Testing

npm test              # Run all tests
npm run test:watch    # Watch mode
npm run test:coverage # Coverage report

Development

npm run dev        # Watch mode
npm run build      # Build for production
npm run rebuild    # Clean and rebuild
npm start          # Run server (stdio)
npm run start:http # Run server (HTTP)
npm run validate   # Typecheck + lint + test

Contributing

Contributions are welcome! Please:

1. Fork the repository

2. Create a feature branch

3. Run tests (npm run validate)

4. Submit a pull request

License

MIT License - see LICENSE for details.

Acknowledgments

• yt-dlp for YouTube extraction

• Anthropic for the Model Context Protocol