Reel Review

Home Page
Content Gallery
The reels we generated!
Our chatbot!

Inspiration

Students spend hours watching hour-long YouTube lectures to learn just 2-3 new concepts and find answers to single questions. Learning, traditionally a social and interactive activity, has become increasingly siloed and lonely in the digital age. We wanted to make education more entertaining and social—not only to save time and improve efficiency, but also to help students enjoy learning for learning's sake. We also understand the frustration of trying to find specific answers in long videos, so we set out to make video content searchable and digestible.

What it Does

Our product takes a YouTube video URL and converts it into a series of short, focused clips. We've implemented a gamified and shareable approach where completing each reel feels like completing a level in a game. Students can share what they've learned on Instagram and LinkedIn, making the learning process social and engaging. We've also integrated a chatbot that answers questions related to the video content using the video's transcript.

How We Built It

Frontend

The website interface is built primarily using React and Tailwind CSS, providing a responsive and modern user experience.

Backend

We use a Python server to handle the core functionality. When a video link is entered, our Python code extracts comprehensive video data from YouTube, including the transcript and video files.

AI Integration

We feed the extracted transcript to a Gemini model that identifies and labels key learning points throughout the lecture. The model outputs structured data including start times, end times, and summaries for each clip segment.

Video Processing

After receiving the AI analysis, we automatically splice the relevant portions of the video and convert them to vertical format optimized for mobile viewing.

Interactive Features

Our chatbot is also powered by Gemini, allowing students to ask specific questions about the video content and receive contextual answers.

Challenges We Ran Into

Ideation: Finding a compelling and feasible project concept took significant effort
Cross-platform compatibility: Ensuring subtitle functionality worked across both Windows and Mac systems (subtitles initially worked on Mac but failed on Windows)
Rate limiting: Getting IP blocked by YouTube during development and testing
API limitations: Reaching the maximum usage limits on the Gemini API during development

Accomplishments We're Proud Of

Successfully implementing an intelligent chatbot that provides contextual answers
Designing intuitive, Duolingo-inspired UI elements that make learning feel like a game
Creating a smooth and reliable link extraction process
Maintaining reasonable work-life balance and getting adequate sleep during development

What We Learned

Development Tools: How to leverage GitHub Copilot effectively for faster coding
AI Integration: Best practices for implementing AI features seamlessly within web applications
Social Features: How to implement social sharing functionality across different platforms
Video Processing: Techniques for automated video editing and format conversion
API Management: Strategies for working within API rate limits and handling service restrictions

Built With

Submitted to

BigRed//Hacks 2025

Created by

I woked on the backend and primarily implemented the link extraction and clip creation methods. It was my first time working with most of the libraries and I learnt a lot.

Anay Garodia
I worked on the front-end, and it was my first time using TailwindCSS. I had never worked in a group larger than two people before, so I learned a lot about how teamwork functions.

Hiroyuki Akiyama
I worked on the frontend with the landing page. I designed the UI in Figma before implementing it. I used many new libraries

JoeCosta7 Costa
I worked on the backend and fullstack integration of the app. I also integrated the AI features of the app with Gemini.

Narayan Topalli