Youtube Knowledge MCP

🚀 YouTube Knowledge MCP

A production-ready Model Context Protocol (MCP) server that transforms YouTube into a queryable knowledge source. It enables searching, fetching details, analyzing transcripts and comments, and powering AI workflows with optional LLMs. Built for Claude Desktop and other MCP clients.

🚀 Quick Start

The YouTube Knowledge MCP is a powerful tool that allows you to leverage YouTube as a knowledge source. With it, you can perform various operations on YouTube data and integrate AI capabilities if needed.

✨ Features

• Fast + quota-aware: It provides fast access to the YouTube API with caching, which is quota-aware.
• Batteries-included tools: Comes with built - in tools for searching videos, getting video details, finding trending videos, and searching channels.
• Optional AI superpowers: If you configure OpenAI or Anthropic API keys, it offers AI - powered features such as generating summaries, topics, chapters, learning paths, analyzing comment intents, and creating knowledge graphs.
• Zero noise: It has minimal configuration requirements, clear logs, and safe default settings.

📦 Installation

Requirements

• Node.js 18+
• YouTube Data API v3 key
• Optional: OpenAI and/or Anthropic API keys for AI tools

Install the package

npm install

📚 Documentation

Configure environment

Create a .env file (or set variables in your MCP client config). You can start from the example:

cp env.example .env

Then set the values in the .env file:

# Required
YOUTUBE_API_KEY=your_youtube_api_key

# Optional AI providers (enables AI tools: analyze_video_content, generate_learning_path, analyze_comment_intents, simplify_video_transcript, generate_video_chapters, generate_knowledge_graph)
OPENAI_API_KEY=your_openai_api_key
ANTHROPIC_API_KEY=your_anthropic_api_key

# Optional tuning
LOG_LEVEL=info
MAX_DAILY_QUOTA=8000
REDIS_URL= # e.g. redis://localhost:6379
REDIS_HOST=
REDIS_PORT=
REDIS_PASSWORD=

An env.example with placeholders is provided. Do not commit your .env file.

Build and run

# Development (watch)
npm run dev

# Production
npm run build
npm start

Connect to Claude Desktop (example)

Add the following configuration to your Claude Desktop with absolute paths:

{
  "mcpServers": {
    "youtube-knowledge": {
      "command": "node",
      "args": ["/absolute/path/to/youtube-knowledge-mcp/build/index.js"],
      "env": {
        "YOUTUBE_API_KEY": "your_youtube_api_key",
        "OPENAI_API_KEY": "optional_openai",
        "ANTHROPIC_API_KEY": "optional_anthropic",
        "LOG_LEVEL": "info"
      }
    }
  }
}

Restart Claude Desktop after editing the configuration.

Available tools

• youtube_search — Search videos with filters
• get_video_details — Video metadata, transcript (best - effort), comments
• get_trending_videos — Most popular by region/category
• search_channels — Channel search with optional stats
• analyze_video_content — AI topics/sentiment/questions/summary/keywords
• generate_learning_path — AI learning path for a topic
• analyze_comment_intents — Classify viewer intents
• simplify_video_transcript — ELI5 - style simplification
• generate_video_chapters — AI chapters with timestamps
• generate_knowledge_graph — Cross - video concept graph

Note: AI tools are available only if an AI provider key is configured.

Quotas and safety

• It enforces a daily quota (default 8000 units) and cost - aware AI usage.
• Logs are sent to stderr (does not break MCP stdio).
• Caching reduces API and token spend; optional Redis is supported.

Troubleshooting

• Missing key: Ensure YOUTUBE_API_KEY is set.
• Quota exceeded: Lower usage, enable caching, or raise MAX_DAILY_QUOTA.
• Claude cannot connect: Verify the absolute path to build/index.js and restart.

📄 License

MIT By Efi Kuta