Refeyn

Inspiration

When previous studying methods failed to work in university classes, the Feynman technique seemed a drastic improvement, but at a steep entrance-barrier in terms of effort. Refeyn makes the Feynman technique more approachable by augmenting your studying experience with AI and an automated review toolkit.

What it does

Feynman allows users to upload a PDF of notes, select individual pages for studying, and create a voice recording of themselves reviewing their selected notes. Refeyn utilizes AI to transcribe the notes PDF into text, uses speech recognition to convert the recording to text, and automatically compares the topics covered in the recording against those in the notes.

How we built it

Refeyn is powered by a React frontend and Django backend. PDF-to-text uses Google Cloud Vision, speech recognition uses the AssemblyAI API, text summarization and semantic similarity comparisons are done using ChatGPT.

Challenges we ran into

Integrating frontend processes with backend proved difficult, as per usual.

Accomplishments that we're proud of

Utilizing a variety of tools for streamlining this widely effective study method.

What we learned

Utilization of various AI APIs, deeper understanding of the Django ecosystem, implementation of theoretical UI design into practical React implementations.