Fetches and cleans transcripts from YouTube videos using yt-dlp with optional Whisper transcription fallback.
This skill empowers Claude to extract text content from any YouTube video by leveraging the powerful yt-dlp utility. It systematically attempts to download high-quality manual subtitles first, falling back to auto-generated captions or even full audio transcription using OpenAI's Whisper model if necessary. Beyond just downloading, it includes a sophisticated post-processing pipeline that deduplicates overlapping lines often found in raw VTT files, providing clean, readable plain-text transcripts perfect for content analysis, summarization, or documentation tasks.
주요 기능
01Advanced deduplication and cleaning for human-readable plain text
02Intelligent fallback to auto-generated captions when manual subs are missing
03Automated subtitle extraction via yt-dlp
04Integrated Whisper transcription for videos without any captions
052 GitHub stars
06Automatic dependency management for yt-dlp and Whisper
사용 사례
01Extracting documentation from video-based software demos
02Creating searchable text archives of video tutorials and guides
03Summarizing long-form video content, lectures, or interviews
