What Inspired Us

Most fitness apps give generic advice. We wanted to build something that knows your body's current state and connects you with the right workout partner.

What We Built

Nexus Twin - a real-time system that:

  • Creates your personal health "twin" with energy/recovery scores
  • Gives AI-powered daily recommendations
  • Matches you with compatible workout partners

How We Built It

  • Frontend: Responsive React with Context state management (mobile + web)
  • Backend: 3 microservices processing health data in real-time
  • Data: Pipeline handling sleep, nutrition, activity, and biometric data
  • Deploy: Docker containers for scalability

Challenges

  • Processing multiple data types simultaneously
  • Making AI recommendations feel personal, not robotic
  • Creating accurate partner matching algorithms

What We Learned

Real-time health data is complex, but when you combine prediction with social connection, people actually stick to their fitness goals.

The result? Your AI health twin that predicts tomorrow's optimal workout and finds you the perfect partner to do it with.

Built With

  • api.meet.no
  • docker
  • fastapi
  • knn
  • multimodal
  • openai
  • openstreetmapapi
  • overpassapi
  • react
  • similarity
Share this project:

Updates