AI-powered injury prevention and performance coaching for extreme sports athletes. Upload a video of your trick or training run and get elite biomechanical analysis with personalized coaching advice — built at IrvineHacks 2026.

1. Athlete Profile — Enter your sport, skill level, body metrics, fatigue level, and injury history
2. Video Upload — Submit a video clip of your trick or training session
3. Pose Detection — MediaPipe extracts joint angles, symmetry, stance width, and movement velocity frame by frame
4. AI Coaching — GPT-5-mini turns the biomechanical data into actionable form corrections, safety warnings, drills, and a conditioning recommendation

Frontend

• React + Vite
• Tailwind CSS

Backend

• FastAPI
• MediaPipe Pose (computer vision)
• OpenCV
• OpenAI API (GPT-5-mini)

redwings-ai/
├── frontend/
│   └── src/
│       ├── components/
│       │   ├── Navbar.jsx
│       │   ├── ProfileForm.jsx
│       │   ├── VideoFeedback.jsx
│       │   └── Results.jsx
│       ├── App.jsx
│       ├── App.css
│       └── index.css
└── backend/
    ├── services/
    │   ├── media_pipe_processing.py
    │   ├── metrics.py
    │   └── llm.py
    ├── api.py
    ├── main.py
    ├── requirements.txt
    └── requirements-dev.txt

• Node.js 18+
• Python 3.10+
• An OpenAI API key
• The MediaPipe pose_landmarker_full.task model file (place it in backend/services/)

You can download the model file here:

https://storage.googleapis.com/mediapipe-models/pose_landmarker/pose_landmarker_full/float16/latest/pose_landmarker_full.task

Create a .env file in the project root:

OPENAI_API_KEY=your_openai_api_key_here

cd backend
pip install -r requirements.txt
pip install -r requirements-dev.txt
fastapi dev src/api.py

The API will be running at http://127.0.0.1:8000.

cd frontend
npm install
npm run dev

The app will be running at http://localhost:5173. Vite proxies all /api requests to the FastAPI backend automatically.

Accepts a video file and athlete profile, returns biomechanical metrics and AI coaching.

Request — multipart/form-data

Response

{
  "profile": { "sport": "Snowboarding", "skill_level": "Intermediate", "..." },
  "metrics": {
    "knee_angle_avg": 132.4,
    "knee_angle_min": 88.1,
    "knee_symmetry_avg": 12.3,
    "hip_angle_avg": 145.2,
    "arm_spread_avg": 0.38,
    "knee_velocity_max": 18.5
  },
  "coaching": {
    "form_corrections": ["...", "...", "..."],
    "safety_warnings": ["...", "..."],
    "drills": ["...", "..."],
    "conditioning": "...",
    "overall_assessment": "..."
  }
}

Built at IrvineHacks 2026 by a team of 4.