We are building just a ton of AI based tools.
How it Works
- Invoke the skill with one command /watch-pdf and the youtube link.
- Moments later you have a nicely formatted PDF and an html page as well.
That's it - and it can't be more simple.
It's based on https://github.com/bradautomates/claude-video that pulls relevant images from the video as well as the transcript and then lets you use Claude to ask questions.
Our add-on adds PDF and HTML pages as training material. Makes for a great combination.
Skills
watch-pdf
Watch a video (URL or local path) and produce a detailed, illustrated training PDF — not just an answer. Runs the watch pipeline (yt-dlp download, ffmpeg frames, transcript), then authors a structured learning document (overview, step-by-step walkthrough with screenshots + captions, key takeaways, glossary) and renders it to PDF (and HTML) with headless Chrome.
watch
Watch a video (URL or local path). Downloads with yt-dlp, extracts auto-scaled frames with ffmpeg, pulls the transcript from captions (or Whisper API fallback), and hands the result to Claude so it can answer questions about what's in the video.
Requirements
yt-dlp, ffmpeg, and a headless Chrome (for watch-pdf).