Learnr is inspired by the pain of trying to copy down text snippets from youtube videos without transcripts.
What it does
Learnr parses text from youtube videos in real time and even offers suggested reading links / definitions for further learning.
How we built it
We built Learnr with electron / react on the frontend and flask / opencv / tesseract / selenium on the back
Challenges we ran into
We initially planned on writing Learnr in a chrome extension but switched to electron after many complications
What we learned
- opencv is impossible to install on linux
- ocr does not play nice with code snippets