Gamifying Self Defense with Machine Learning
Kravity is a mobile-first AR platform built to make self-defense training fun, accessible, and engaging.
Developed at HopHacks Fall 2025 by Ethan Wang, Alyssa Xu, and Ankur Senapati, Kravity transforms physical motion into an interactive self-defense experience through real-time AR tracking, gamification, and personalized feedback.
Traditional self-defense training can feel repetitive, intimidating, and inaccessible. We wanted to reimagine the process — blending gaming mechanics, augmented reality, and motion tracking to make training immersive, safe, and fun.
- AR Motion Tracking – Uses camera-based body tracking to detect and grade defensive stances.
- Gravity-Based Challenges – Real-time simulations force quick reactions under pressure.
- Gamification Layer – XP, streaks, leaderboards, and progress tracking to encourage consistency.
- Adaptive Feedback – Personalized guidance on form, timing, and accuracy.
- Languages: JavaScript, TypeScript
- Frameworks: React, Next.js
- APIs & Tools: MediaPipe, TensorFlow.js
- Platforms: Web + Mobile (React Native)
- Cloud Services: Firebase (Auth, Firestore, Hosting)
- Design: Figma, Canva
Check out our full pitch deck for visuals, problem framing, and demo flow:
👉 Kravity Pitch Deck
-
Ethan Wang
-
Alyssa Xu
-
Ankur Senapati
Built at HopHacks Fall 2025
Devpost Project Page
- Multiplayer dojo training mode
- More motion-detection modules
- Partnerships with gyms & martial arts studios
- AI-powered personalized training plans
Clone the repo and install dependencies: