Connects to YouTube Data API v3 to pull video metadata, transcripts, channel stats, search results, trending content, and comments. Ships with aggressive token optimization that strips 64-87% of the payload bloat from raw API responses, which matters when you're feeding video data into context windows. Includes optional MongoDB caching to protect your daily API quota from agent retry loops. Works out of the box without an API key if you only need transcripts, but you'll want to configure credentials to unlock search, analytics, and channel tools. Built with 97% test coverage and strict typing, so it won't randomly break your workflows. Reach for this when you need YouTube data in AI applications without burning tokens or quota on redundant JSON.
claude mcp add --transport stdio ai.smithery-kirbah-mcp-youtube uvx kirbah-mcp-youtube