yt-dlp MCP Server

Example Use Cases

• 1

  Search YouTube programmatically and return paginated, JSON- or Markdown-formatted results for downstream processing or presentation.

• 2

  Extract detailed video metadata (views, likes, tags, upload date, formats) to inform recommendations, indexing, or analysis workflows.

• 3

  Download videos or audio, trim content, or retrieve subtitles/transcripts (cleaned text or VTT) for summarization, translation, or content ingestion.

Description

This MCP server exposes yt-dlp functionality to AI agents (Claude, Dive, etc.), allowing them to search videos, fetch comprehensive metadata, download video/audio, and retrieve subtitles or cleaned transcripts. It includes tools with Zod validation, character limits for LLM safety, and convenient response formats (JSON/Markdown) for programmatic or human consumption. Installation is via npm or running the package as an MCP server in Dive Desktop, and it is designed to be privacy-friendly with no tracking.