Inspiration

We have all struggled in class, whether the teacher was unclear, or you're having trouble understanding a topic. If you don't understand the material, it will lead to failed exams and unnecessary stress. But now with our product, Revise, you can change that.

What it does

All you need to do is upload the audio recording of your class(.wav file), and our AI models will give you a summary of what you have learned in class and it will give you a few questions to test your understanding.

How we built it

I used streamlit to built the website for our app and we used NLTK, pytorch, transformers and numpy for our AI Models. In addition, we used pydub to process our audio files and speech recognition to convert the audio files to text.

Challenges we ran into

I had trouble processing large audio files as we could not install one of the necessary libraries. Originally we used Django to make our python app but we ran into issues deploying our AI models to the website. So I had to abandon our Django app, and start all over again with streamlit as our webapp framework.

Accomplishments that we're proud of

I was proud to learn new technologies to make this this web app! I learned multiple new technologies like Natural Language Processing and StreamLit and I am also proud of the end result of our app! Seeing all the hard work put in being paid off is an amazing experience! I am proud that I was able to still manage to finish our app despite us switching web dev frameworks in the middle and starting all over again.

What we learned

I learned new technologies like NLP and Streamlit and many more. This entire hackathon was a great learning experience for all of us!I learned so many new things and we also improved our communication skills as they were essential for success.

What's next for Revise.ai

We need to implement a recording feature so you can record your meetings/classes and upload them directly to our AI. We also wanted to make our AI models and our entire website more efficient and faster as well as more accurate with answers in the future.

Built With

Share this project:

Updates