Class Buddy

This is the dashboard

Inspiration

As students, we constantly missed deadlines buried in endless class group messages. Whether it was an assignment update or a class reschedule, important info often got lost in the chaos. Talking to classmates, we realized we weren’t alone. Everyone was struggling to keep track. So we built what we wished we had—ClassBuddy.

What it does

ClassBuddy is an AI-powered academic assistant that:

Monitors class group messages
Extracts and summarizes important information like assignments, exams, class changes
Notifies students before deadlines approach
Generates voice and video digests using ElevenLabs and Tavus

It helps students stay on top of academic tasks without manually checking chat history.

How we built it

We built ClassBuddy using:

React + TypeScript + Tailwind CSS for a responsive and intuitive UI
Claude and Gemini APIs to summarize chat messages into structured task lists
ElevenLabs API to convert summaries into natural voice digests
Tavus API to generate personalized video summaries
Custom message simulation and monitoring loop to test in a live-like setting

Challenges we ran into

Designing effective prompts to get structured outputs from LLMs like Claude and Gemini
Handling latency in voice and video generation APIs
Ensuring messages weren’t processed multiple times
Presenting data in a clean, digestible way while maintaining functionality

Accomplishments that we're proud of

Real-time chat monitoring with automated AI summarization
Seamless voice and video digest generation for academic tasks
A sleek and easy-to-use interface tailored for student workflows
Built something genuinely useful that solves a real-world problem

What we learned

Prompt engineering is critical for AI reliability
Students want proactive tools—not just passive ones
Combining multiple AI modalities (text, voice, video) adds significant user value
Simulating real-world usage early helps catch edge cases