It uses youtube-transcript-api to get the captions from YouTube. Then it separates the captions into 5 chunks and generates questions for each chunk. It uses a deployed Hugging Face model to generate the questions.
It uses Vercel Serverless Function to run the flask server. Hopefully in under 10 seconds.
pip install -r requirements.txtpython app.py/POST- body: {"youtubeVideoId": "g0amdIcZt5I"}
anyGET