MCP-Youtube

(c)2026 - Dr. Edwin A. Hernandez

MEVIA Project

A Modern Context Protocol (MCP) server for YouTube that retrieves and processes video content.a

Features

• Retrieve YouTube video metadata (title, description, thumbnail)
• Extract and process video transcripts
• Generate video summaries
• Find similar videos
• Translate content to different languages
• Support for multiple transport methods: TCP, STDIO, and SSE

Prerequisites

• Python 3.9+
• Google Cloud Project with YouTube Data API v3 enabled
• API key for YouTube Data API

Installation

1. Clone the repository:

git clone https://github.com/yourusername/MCP-Youtube.git
cd MCP-Youtube

2. Install dependencies:

pip install -r requirements.txt

3. Create a .env_yt file with your YouTube API key:

YOUTUBE_API_KEY=your_api_key_here
OPENAI_API_KEY=your_openai_api_key_here  # For translation and summarization

Note: The .env_yt file is already listed in .gitignore to protect your API keys. Do NOT commit your .env_yt file to version control.

Usage

As an MCP Server

MCP-Youtube implements the Modern Context Protocol and can be used with LLMs that support MCP.

When an LLM encounters a YouTube URL (e.g., http://youtube.com/watch?v=VIDEO_ID), MCP-Youtube will:

1. Retrieve video metadata (title, description, thumbnail URL)
2. Extract the transcript
3. Generate a summary
4. Provide similar video recommendations
5. Support translation of content

Transport Methods

MCP-Youtube supports multiple transport methods:

1. TCP Transport: Connect via TCP socket
2. STDIO Transport: Use standard input/output for communication
3. SSE Transport: Server-Sent Events for streaming communication

API Endpoints

• /youtube/video/{video_id}: Get video metadata
• /youtube/transcript/{video_id}: Get video transcript
• /youtube/summary/{video_id}: Get video summary
• /youtube/similar/{video_id}: Get similar videos
• /youtube/translate: Translate content

Configuration

Environment variables:

• YOUTUBE_API_KEY: Your YouTube Data API key (required)
• OPENAI_API_KEY: Your OpenAI API key for translation/summarization (optional)
• MCP_PORT: Port for TCP transport (default: 3000)
• LOG_LEVEL: Logging level (default: INFO)

Development

1. Install development dependencies:

pip install -r requirements-dev.txt

2. Run tests:

pytest tests/

3. Run the server:

YOUTUBE_API_KEY=#####. python3 -m mcp_youtube.mcp_server --transport http --port 9091

License

MIT License