Nexus

Make learning any class fun and simple with your friends!

Comment 1

Inspiration

  • For us (and many others), studying is often a troubling experience. Traditional study methods can feel monotonous and isolating, and it becomes especially difficult if you are starting from scratch. While online platforms like Khan Academy exist to help students learn, they are often limited to predefined topics and even paywalls. They also lack the social element that makes learning engaging.
  • We envisioned Nexus as a way to put the power in the hands of students. What if students could create their own "Algebra Nation" for any subject they wanted to study? What if they could generate interactive lessons, quizzes, and study materials on-demand for any topic, whether it's art history or quantum physics.
  • And on top of that, we decided to embed all of this functionality within a real-time messaging app so you can learn with your friends, solving the issue of studying feeling isolating and boring.
  • By allowing students to create their own modular learning experiences with their friends in real time, Nexus aims to transform studying from a solitary chore into an interactive, social experience where students can learn together under the assistance of an adaptive intelligent AI agent.

What it does

  • Nexus is an intelligent study platform that combines real-time collaboration with AI assistance. By embedding AI assistants within real-time chats with users, it creates a dynamic learning environment where students can receive instant help, generate study material, take interactive quizzes, and collaborate with peers while maintaining context aware conversations that adapt to their learning needs and progress.

How we built it

  • frontend using react with vite
  • node.js/express server with MongoDB Atlas for data and ai memory storage
  • socket.io for live chats and updates
  • google's gemini 2.0 flash ai for:
    • processing study material
    • general lesson plans
    • creating interactive quizzes
    • providing adaptive guidance to users
  • authentication using custom JWT based auth system
  • multer for handling file uploads

Challenges we ran into

  • embedding AI assistants within real-time chat messages
  • managing real-time state synchronization between multiple users
  • handling concurrent file uploads and processing
  • maintaining context across different AI conversations and persistent memory within AI
  • ensuring smooth integration between the chat system and AI responses
  • optimizing performance with large amounts of study material

Accomplishments that we're proud of

  • seamless integration of an ai assistant within a real-time messaging application/study tool
  • communication and context passing between different instances of ai within different chats
  • persistent memory and adaptive logic tracking each user's progress
  • real-time collaborative features that work smoothly across multiple users
  • intelligent study material processing and lesson plan generation
  • interactive quiz system with progress tracking
  • clean and intuitive user interface

What we learned

  • how to maintain conversation context in AI models that don't support persistent memory.
  • how to effective handle real-time updates in a multi-user environment
  • techniques for optimizing AI responses with efficient prompt engineering

What's next for Nexus

We have many plans that we weren't able to implement within the time frame of this hackathon. First and foremost we want to add voice calls, as studying is usually done with your peers with some kind of voice communication. Voice integration with the AI assistant comes naturally as well. We also want to implement the Google Calendar, as the Gemini API is able to edit the Google Calendar by itself. We would want to take all user's course schedules and construct detailed study plans on their personal calendars using Gemini's reasoning.

Built With

Share this project:

Updates