ReelyPolly

Scrolling? Do your quizzes first!
Comments are sneakily replaced with lecture content. You can't run!

Inspiration

Students already scroll YouTube Shorts during lectures.

Every solution we’ve seen tries to fix this, attention tracking, engagement tools, active learning platforms. We decided to ask the forbidden question: What if we made distraction the feature instead of the bug?

ReelyPolly was inspired by the modern attention economy, shrinking attention spans, and the uncomfortable reality that short-form content often wins over traditional lectures. Instead of fighting it, we leaned all the way in.

What it does

ReelyPolly makes lectures more interesting in all the wrong ways.

It transforms lectures into a YouTube Shorts–style experience, where students can scroll endlessly while “attending” class. When lecturers launch pop quizzes, they appear seamlessly inside the Shorts feed, no tab switching, no attention required.

Students can:

Answer quizzes directly within YouTube
Skip questions entirely and let AI automatically answer quizzes for them.
Automatic live transcription even handles the peskiest of quizes, those that can only be answered by actively listening to the lecture.

The result: maximum reels scrolled, minimum lecture absorbed.

How we built it

ReelyPolly is built as a real-time system designed to minimize friction between distraction and “participation.”

Frontend

The frontend is a React-powered Chrome extension that lives directly in the user’s browser. By embedding itself into YouTube, it transforms the Shorts feed into an interactive lecture experience.

The extension:

Renders lecture content and pop quizzes natively within the Shorts interface
Detects incoming quizzes in real time
Allows students to answer, skip, or delegate responses to AI without leaving YouTube

This ensures uninterrupted scrolling at all times.

Backend

The backend is implemented in Python and orchestrates real-time audio processing, quiz handling, and AI interaction.

Key components include:

Whisper for live lecture transcription
LangChain to manage prompt flows, Retrieval Augmented Generation (RAG) and quiz-answering logic
OpenAI API to generate AI-powered quiz responses
WebSockets for low-latency, real-time communication between lectures and the Chrome extension

Together, these components allow quizzes to appear instantly inside the feed and be answered automatically when students choose not to engage.

Challenges we ran into

Loads of debugging! With the large number of components, it was hard to integrate everything seamlessly.

Accomplishments that we're proud of

We're proud of prioritising student mental health in an era plagued by abuse by lecturers. In all seriousness, we are genuinely proud of the number of features we came up with and integrated into our product in such a short amount of time.