- ! Create a new streamlit project for youtube to mp3 web client
Youtube to AI Summary
- Create POC with StreamLit and backend with FastAPI.
- send to Nate and post on discord servers for feedback
- if valid idea
- Create React + Tailwind frontend for better aesthetics
Ideas
-
review and look at old fastAPI project
- understand how it works again
- integrate with pytest
- deploy it as an API on that one API website where you make money after X many requests.
-
@ Once I am finished with the basic development version of the obsidian plugin, and the website: add the ability to scrap the youtube comments for insights too.
- maybe filter importance by the number of likes of reply's to the comment.
-
@ as well as scrape the description of the youtube video for links/resources that can be added to the summary.
-
! also re-work the youtube to AI summary script to have a "--detailed" flag
that allows for an in-depth breakdown of the transcript.
eventually just build like a gradio project to share on hugging face, that is youtube client where you can type a topic or an idea
- searches for the idea
- grabs top 5 videos
- makes and compares the youtube summaries
- use ML or DL to judge how good the summary is
- returns a write-up about the search query and gives recommendations for the best videos to watch for the most context.
Update and get the fastapi project running again on Ubuntu
postman
Docker
PostgreSQL
Alembic
Pydantic
Etc
Breakdown the just code in obsidian so the concepts stick and can be referenced,
Finish the tutorial and add unit testing and stuff with pytest or unit test
instead of restarting the tutorial or continuing, because it is outdated. follow along with the fastapi documentation and try to reimplement all the features this project has
Use videos as a guide but use documentation first
Youtube Resources
FastAPI Documentation
Done
-
- branch name is: update-project