Vantage Lifts ⚡️ Proof of Sweat Protocol

Targeting the Polychrome Mosaic Track at TartanHacks 2026 > Bridging Preventative Healthcare & Decentralized Finance through Computer Vision.

🚀 The Elevator Pitch

Vantage is the world's first "Proof of Sweat" ecosystem. While traditional fitness apps rely on manual (and easily faked) inputs, Vantage uses real-time Computer Vision and Velocity Loss Tracking (VLT) to cryptographically verify physical exertion.

By turning biomechanical data into a verifiable asset, we unlock a new layer of "Active Finance":

  • For Users: Stake credits on your own workout performance.
  • For Health: AI-driven form correction that prevents injury before it happens.

🛠 The Tech Stack

  • Frontend: React Native (Expo)
  • Backend: Python (FastAPI + WebSockets)
  • AI Engine: YOLOv26n-pose (Pose Estimation via Ultralytics)
  • Real-time Protocol: JSON over WebSockets (Avg latency < 100ms)

📸 How to Run the Demo (Local Setup)

For the MVP, the backend runs locally on your laptop and streams data to your phone via local WiFi.

Prerequisites

  1. Node.js & npm
  2. Python 3.9+
  3. Expo Go App installed on your iOS/Android device.
  4. Important: Your laptop and phone MUST be on the same WiFi network (or use a mobile hotspot).

Step 1: Fire Up the AI Backend (Python)

  1. Navigate to the backend folder: bash cd backend
  2. Create and activate a virtual environment (optional but recommended): bash python3 -m venv venv source venv/bin/activate # Windows: venv\Scripts\activate
  3. Install dependencies: bash pip install fastapi uvicorn ultralytics opencv-python websockets
  4. CRITICAL: Find your Laptop's Local IP Address.
    • Mac: Run ipconfig getifaddr en0 in terminal.
    • Windows: Run ipconfig and look for "IPv4 Address".
    • Keep this number handy (e.g., 192.168.1.5).
  5. Start the server: bash python3 server.py

Step 2: Connect the Frontend (React Native)

  1. Open frontend/app/index.js.
  2. Find the line defining SERVER_URL: javascript // REPLACE THIS with the IP you found in Step 1 const SERVER_URL = "ws://192.168.1.5:8000/ws/analyze";
  3. Navigate to the frontend folder in a new terminal: bash cd frontend
  4. Install dependencies: bash npx expo install expo-camera expo-gl
  5. Start the Expo server: bash npx expo start

Step 3: Start Pumping Iron!

  1. Scan the QR code displayed in your terminal with your phone (or the Expo Go app).
  2. Grant Camera Permissions when prompted.
  3. Point your camera at a teammate doing a bicep curl or squat.
  4. Watch the Angle, Rep Count, and Wager update in real-time!

🛡️ Privacy by Design

Vantage is built on the principle that your biometric data belongs to you alone. We use a Privacy-First architecture to ensure your workout remains secure and anonymous.

  • Edge AI (On-Device Processing): 100% of the AI pose estimation happens locally on your smartphone. Your camera feed is never uploaded to the cloud or stored on a server.

  • Zero-Recording Policy: The app does not record or save raw video. It interprets motion as a mathematical stream of $(x, y, z)$ coordinates and discards visual data in real-time.

  • The "Bystander Shield": To respect the privacy of others in the gym, Vantage uses semantic segmentation to apply a real-time blur to the background, ensuring only the primary user is visible to the AI.

  • AR Skeletal Masking: Instead of video, the app projects an Augmented Reality "stick figure" over your form. You interact with biomechanical data, not sensitive footage.

  • Data Sovereignty: All "Workout Scores" are stored in a secure local vault. Data is only shared with friends or for "Proof of Sweat" wagering when you explicitly authorize it to share the info with your friends.

⚠️ Troubleshooting

  • "Network Request Failed" / No Connection:
    • Ensure your phone and laptop are on the same WiFi.
    • If on University/Public WiFi: Create a Personal Hotspot from your phone and connect your laptop to it.
    • Try running Expo with a tunnel: npx expo start --tunnel.
  • Laggy Video:
    • We deliberately downscale images for the hackathon demo to ensure speed over quality.
    • Check the setInterval in App.js. If it's too fast (e.g., 50ms), try increasing it to 150ms.

🔮 Roadmap

  • Smart Contract Integration: Move "Wager" logic to an on-chain escrow contract.
  • Tele-PT: Allow remote coaches to see the "Skeletal Overlay" in real-time.
  • Hardware Integration: Apple Watch Heart Rate correlation for "CNS Fatigue" scores.

Built with 💻 and ☕️ for TartanHacks 2026.

Share this project:

Updates