A solid YouTube transcript extractor that actually works reliably. Pulls subtitles and captions directly from YouTube's API with automatic fallback to yt-dlp when blocked. Handles multiple languages, auto-generated and manual captions, and caches everything locally in organized folders. Outputs clean markdown with timestamps, SRT files for video editors, or raw data for further processing. The chapter parsing from video descriptions and speaker identification setup are nice touches. Smart caching means you can reformat the same video multiple times without hitting YouTube again. Does exactly what it says without unnecessary complexity.
claude skill add jimliu/baoyu-skills:baoyu-youtube-transcript