FocusFlow: Turning Educational Videos into Personalized Notes in Real Time

Inspiration

In a world overflowing with educational videos—from Khan Academy to YouTube to Collegboard—students often lose focus while trying to take notes and understand the context effectively. Our team experienced this first-hand during long study sessions. We imagined a tool that could streamline learning by capturing the essence of a video while maintaining engagement and comprehension. Thus, FocusFlow was born: a companion that keeps you locked in.

What it does

FocusFlow transforms any educational video into structured, concise, and personalized notes in real time. It extracts the video transcript, identifies key concepts, formulas, and examples using natural language processing (NLP), and formats them into easy-to-read summaries. Users can paste a link from CollegeBoard, YouTube, or Khan Academy, and FocusFlow generates intelligent notes.

How we built it

We developed FocusFlow as a browser extension built with HTML, CSS, TypeScript, and Python. The backend uses a Python Flask server integrated with Gemini for transcript analysis and summarization. We used YouTube’s transcript API and web-scraping for text extraction, followed by a structured pipeline that classifies information into headers, bullet points, and keywords.

Challenges we ran into

We faced a multitude of challenges. We initially started with the project being a Chrome extension. However, our lack of expertise in making one made us loose half the day working on it. At 2 pm, we concluded that we should transition to a website. Within the Website, our GUI, Webpages all work however the API for Google Gemini occasionally posed issues in our project. Overall, it was challenging but allong with these conflicts we had successes.

Accomplishments that we're proud of

We successfully built a fully functional prototype capable of generating usable notes. The notes that feel natural; they feel as if a human wrote them. We also designed a minimalist interface that enhances focus rather than distracting from it. Hence, it is very true to our project’s name.

What we learned

We learned how to integrate multiple APIs efficiently. We also learned how to handle complex natural language data. Most importantly, we learned that simplicity and clarity in design are just as important as powerful AI capabilities.

What's next for FocusFlow

We plan to expand FocusFlow into a full-featured platform with user accounts, personalized study analytics, and support in multiple different languages. Future versions will include spaced-repetition learning integration, for this is a very effective and widely used, and popularized learning method.