Hacker News new | ask | show | jobs
by XCSme 15 days ago
I was excited to use it, as I really wanted something like this, then I realized it needs AI/Claude (?)

That sounds like it can get quite costly. Probably there are ways to do it without AI, I would rather manually annotate the tab area with a visual editor.

1 comments

How would you do this without SOME sort of vision/audio model? Are you saying just not an LLM?
Manually select a rectangle on a video frame, then do basic computer vision to detect notes, or even a simple image processing algorithm to find the lines and notes.