StudyRot

Inspiration

Studying is a heavy and boring process. However, we social media like X and TikTok caught us in an endless loop of scrolling. What if we could make studying as addicting as scrolling?

What it does

StudyRot transforms heavy academic materials into an engaging threads. It acts as a social-media-inspired mindmap:

Breadth vs. Depth: The main thread provides a high-level overview of the topic (breadth). If a student understands it, they keep scrolling. If they are confused, they click into the "replies" of that specific tweet to drill down into detailed explanations (depth).
Automated Generation: Users upload a syllabus or PDF, and StudyRot generates a progressive, story-driven thread.
Engagement via Misconceptions: Instead of just stating facts, the AI deliberately incorporates common misconceptions or "ragebait" into the threads to spark critical thinking and keep users actively engaged in the narrative.

How we built it

Document Processing: Users upload files (PPTs, PDFs), which are processed using OCR and LLMs to extract and summarize the core content.
Vectorization: The backend breaks this content into logical chunks, generates embeddings, and stores them in a vector database.
RAG Engine: When a thread is generated, we use Retrieval-Augmented Generation (LLM + embeddings) querying the knowledge base ID. This ensures the "thread" and "replies" are factually accurate, contextually relevant, and chronologically sound.
Frontend: We designed a UI that mimics the frictionless, intuitive scrolling mechanics of popular social media platforms.

Challenges we ran into

New Tech Stack: Many of us learned some of the tech stacks (e.g., TCPR, Drizzle) for the first time. Although challenging, it significantly expanded our skillset within this short development time.
Time Management: 24 hours to create a full stack web application was extremely hard, having to plan out and design the system from scratch, and having to do much pull requests which needed to be reviewed made us underestimate the amount of time needed.
Over Engineering: Over engineering became a problem as we were all trying to maximize the product features, which led to much confusion when we were trying to integrate all the individually designed components.

Accomplishments that we're proud of

Successfully disguising a complex educational tool (mindmapping) inside an interface that feels exactly like scrolling through social media.
Building a fully functional RAG pipeline from scratch that accurately grounds the AI's threads in the user's specific uploaded syllabus.
Creating a learning tool that actually caters to slow learners and students trying to catch up, giving them the agency to choose their own depth of learning without feeling overwhelmed.

What we learned

UX is just as important as the AI: We learned how deeply UI formatting affects reading comprehension. Breaking text into threads drastically reduces cognitive load.
Advanced RAG Implementation: We gained deep, hands-on experience with chunking strategies, vector databases, and Retrieval-Augmented Generation.
Synchronizing with Teammates: Working together for a solution within such a short time period made us realize how important it was to have synergy with our teammates, especially in synchronizing the development process and ensuring it was as smooth as possible.
Getting familiar with the Tech Stack Beforehand: Considering the time limit, it would be ideal for the team to come together before the hackathon to really discuss, agree, and familiarize the tech stack used during development to minimize the chances of hitting bumps during development

What's next for StudyRot

Gamification & Leaderboards: Implementing leaderboards with connected people to motivate students to improve.
Interactive Checkpoints: Adding swipeable flashcards (swipe right for material understood, left to select material for further testing) integrated directly into the feed to test knowledge before the user can scroll further.
Multimedia Integration: Automatically generating AI images, short videos, or background music to attach to threads for an immersive, multi-sensory learning experience.