YouTube MCP Server

MCP (Model Context Protocol) server to interact with YouTube: search videos, get transcripts, find trending content, and explore AI communities.

Features

• Search videos by keyword
• Get trending videos by region and category
• Fetch video transcripts/subtitles (no API key needed for transcripts)
• Get detailed video metadata
• Look up channel info and latest uploads
• Read video comments
• Discover and rank AI-focused communities by subscribers

Prerequisites

• YouTube Data API v3 key — get one at https://console.cloud.google.com/apis/credentials
• Enable the YouTube Data API v3 in your Google Cloud project

Quick Start

cp .env.example .env
# Edit .env with your YOUTUBE_API_KEY

pip install -r requirements.txt
python server.py

Tools

Tool	Description	Arguments
search_videos	Search videos by keyword	query (string, required), max_results (int, default 10, max 50)
get_trending_videos	Trending in a region	region_code (string, default BR), category_id (string, optional)
get_video_details	Full video metadata	video_id (string, required)
get_video_transcript	Video transcript/subtitles	video_id (string, required), lang (string, default pt)
get_channel_info	Channel profile and stats	channel_id (string, required)
get_channel_videos	Latest videos from channel	channel_id (string, required), max_results (int, default 10)
get_video_comments	Comments sorted by relevance	video_id (string, required), max_results (int, default 20, max 100)
search_ai_communities	Find AI communities, sorted by subscribers	query (string, default 'artificial intelligence'), max_results (int, default 10)

Register with opencode

{
  "mcpServers": {
    "youtube": {
      "command": "python3",
      "args": ["/caminho/para/youtube-mcp/server.py"],
      "env": {
        "YOUTUBE_API_KEY": "sua_chave"
      }
    }
  }
}

How it works

Uses the YouTube Data API v3 for metadata, search, and comments, and youtube-transcript-api for subtitles (no extra auth needed).

Docker

This is a stdio server, so the container must keep stdin open (-i):

docker build -t youtube-mcp .
docker run -i --rm -e YOUTUBE_API_KEY="sua_chave" youtube-mcp

A HEALTHCHECK validates the MCP handshake on startup.

Tests

python -m pytest . -v