I came up with this idea because I was enrolled in an online course and it was all videos. I spend all day long on the computer for work, so the last thing I want to do after work is spend more time staring at a screen. I wanted something that could take those videos and turn them into a readable format.
The most difficult part was detecting what frames in the video are slides, and extracting from those what is useful content to add to the document. I didn't want to just stick a screengrab of the slides in the text, I wanted it to read like an actual book, so it uses an AI model to take from the slides what is useful and ignores what is just repeated in the transcript.
Here is an output sample from an NIH video: https://lecturetobook.com/sample/sleep-book/book.html
Feedback welcome! This is my first side-project that has made it to launch.