What problem does it solve?
This Skill automates obtaining transcripts, subtitles, and audio extraction from YouTube and other yt-dlp supported sites, saving transcripts to file by default. Use when you need to convert video content into readable, searchable text for documentation, indexing, or archival purposes.
Core Features & Use Cases
- Transcript extraction: Fetch transcripts for YouTube videos via youtube-transcript-plus when available, with language selection.
- Fallback + Subtitles: Fall back to yt-dlp subtitles when direct transcripts are unavailable and also support downloading subtitles or audio.
- Export & Metadata: Save transcripts as Markdown files with video metadata frontmatter including title, date, id, uploader, views, and duration; supports saving to file or printing to stdout.
- Use Case: Create searchable documentation from a lecture video by generating a clean, paragraph transcript and attaching metadata for indexing.
Quick Start
Install dependencies with ./install.sh, then run the script to fetch a transcript:
- Transcript: ./scripts/vtd.js transcript --url 'https://www.youtube.com/watch?v=...' --lang en
- Subtitles: ./scripts/vtd.js subs --url 'https://www.youtube.com/watch?v=...' --lang en
- Download: ./scripts/vtd.js download --url 'https://www.youtube.com/watch?v=...' --output-dir ~/Downloads