Unlock your athletic potential through AI-powered motion analysis and intelligent feedback.

Dynamis is a real-time sports training platform that captures an athlete's movement, analyzes key biomechanical features, and delivers AI-generated coaching feedback to optimize performance, form, and injury prevention.

Leveraging computer vision, real-time data processing, and large language models (LLMs), Dynamis helps athletes train smarter — unlocking the full power of their motion.

• 📷 Real-Time Pose Tracking — Capture body landmarks using a standard camera.
• 📊 Movement Feature Extraction — Analyze joint angles, stability, and motion dynamics.
• 🧠 AI Feedback Generation — Use LLMs to interpret performance and recommend improvements.
• 🔊 Dynamic Speech Coaching — Deliver coaching advice through AI-generated audio.
• ⚡ Multi-Sport Flexibility — Built to adapt across different sports and training activities.

• Frontend (Vite + React):

  • Captures video stream and extracts body pose landmarks.
  • Sends batches of movement frames to backend for analysis.
  • Plays AI-generated audio feedback to the user.

• Backend (FastAPI):

  • Processes real-time pose data and extracts key athletic features.
  • Summarizes movement metrics (e.g., elbow extension, wrist snap, hip stability).
  • Sends feature summaries to LLM for evaluation.
  • Returns a coaching feedback message to frontend.

• AI Layer:

  • Google Gemini interprets biomechanical summaries.
  • OpenAI TTS converts feedback into dynamic spoken advice.

git clone https://github.com/beritcheema/dynamis.git
cd dynamis

cd backend
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

Create a .env file inside /backend/:

GEMINI_API_KEY=your_google_gemini_api_key
OPENAI_API_KEY=your_openai_tts_api_key

Start the backend server:

uvicorn main:app --reload --port 8000

cd ../
npm install
npm run dev

Frontend will be available at http://localhost:5173.

Train Smarter. Move Better. Unlock Your Potential.