MyCoachAI revolutionizes fitness training by transforming your smartphone into an intelligent personal trainer.
MyCoachAI revolutionizes fitness training by transforming your smartphone into an intelligent personal trainer. Using advanced computer vision and machine learning, it provides real-time form analysis, instant feedback, and personalized coaching for your workouts.
🔥 No Equipment • 📱 Just Your Phone • 🤖 AI-Powered • 🔒 100% Private
- 🎯 Real-Time Form Analysis - AI tracks 17 key body points for perfect technique
- 🗣️ Instant Audio Coaching - Get voice feedback without looking at your screen
- 📊 Smart Rep Counting - Only quality movements count toward your goals
- 🧠 Adaptive Programming - Workouts that evolve based on your progress
- 📈 Progress Analytics - Visual insights into your form improvements
- 🔒 Privacy First - All AI processing happens on your device
- Python 3.11+ (for backend development)
- Node.js 18+ (for mobile app development)
- Git (for version control)
# Navigate to backend
cd backend
# Install Python dependencies
pip3 install -r requirements.txt
# Start the API server
uvicorn app.main:app --host 0.0.0.0 --port 8000 --reload
2. Test the API
Visit http://localhost:8000/docs to explore the interactive API documentation.
Try the pose analysis endpoint:
curl -X POST "http://localhost:8000/api/v1/pose/analyze" \
-H "Content-Type: application/json" \
-d '{
"exercise_type": "squat",
"keypoints": [{"x": 100, "y": 200, "confidence": 0.9}],
"timestamp": 1234567890
}'
🏗️ Project Structure
MyCoachAI/
├── 📱 apps/
│ ├── mobile/ # React Native app (iOS/Android)
│ └── web/ # Next.js landing page
├── 🔧 backend/
│ ├── app/
│ │ ├── api/v1/ # FastAPI routes
│ │ │ ├── pose.py # Pose analysis endpoints
│ │ │ └── workouts.py # Workout endpoints
│ │ ├── services/ # Business logic
│ │ │ └── pose_analyzer.py # AI form analysis
│ │ └── main.py # FastAPI application
│ └── requirements.txt # Python dependencies
├── 🏢 infrastructure/
│ ├── docker/ # Container configurations
│ └── terraform/ # AWS infrastructure as code
├── 📦 packages/
│ └── ui/ # Shared UI components
└── 📚 docs/ # Documentation
Tech Stack
Layer Technology Purpose
Mobile React Native + Expo Cross-platform mobile app
AI/ML TensorFlow.js + MoveNet On-device pose estimation
Backend FastAPI + Python REST API and business logic
Infrastructure AWS + Docker Scalable cloud deployment
🎯 Supported Exercises
✅ Currently Available
Exercise Difficulty Form Checks
Squats Beginner Depth, knee alignment, posture
Push-ups Beginner Body alignment, range of motion
🔄 Coming Soon
Lunges - Balance and form analysis
Deadlifts - Hip hinge pattern recognition
Planks - Core stability tracking
Pull-ups - Grip and pulling mechanics
🔬 How It Works
1. Pose Detection
Uses MoveNet Lightning model to detect 17 key body points:
text
nose, left_eye, right_eye, left_ear, right_ear,
left_shoulder, right_shoulder, left_elbow, right_elbow,
left_wrist, right_wrist, left_hip, right_hip,
left_knee, right_knee, left_ankle, right_ankle
2. Form Analysis
python
def analyze_squat(keypoints):
# Check depth: hip below knee level
if hip_y > knee_y:
return "Great depth! 💪"
else:
return "Go deeper - hip should be below knee level"
3. Real-Time Feedback
Form Score: 0-100 based on technique quality
Audio Cues: "Keep your back straight", "Great form!"
Visual Analysis: Detailed movement breakdown
🛠️ Development
GitHub Codespaces (Recommended)
Open in GitHub Codespaces
Why Codespaces?
✅ All dependencies pre-installed
✅ Port forwarding for API testing
✅ VS Code with extensions
✅ Zero local setup required
API Endpoints
Endpoint Method Description
/ GET API status and info
/health GET Health check
/docs GET Interactive API documentation
/api/v1/pose/analyze POST Analyze pose data
/api/v1/pose/exercises GET Supported exercises
/api/v1/workouts/next GET Get personalized workout
📱 Mobile App (Coming Soon)
Features in Development
Camera Integration - Use phone camera for pose detection
Real-time Overlay - Visual feedback during workouts
Workout Programs - Curated routines for all fitness levels
Progress Tracking - See your improvement over time
🤝 Contributing
We welcome contributions from developers, fitness experts, and AI researchers!
Ways to Contribute
🐛 Report Bugs - Help us improve stability
💡 Suggest Features - Share your ideas for new exercises
🔧 Submit PRs - Contribute code improvements
📝 Improve Docs - Help others understand the project
🚀 Roadmap
Phase 1: MVP (Current)
✅ Real-time pose detection API
✅ Basic exercise analysis (squats, push-ups)
✅ Form scoring algorithm
🔄 Mobile app development
Phase 2: Enhanced Features
📱 iOS/Android app release
🗣️ Real-time audio coaching
📊 Progress tracking and analytics
🏋️ Additional exercises (deadlifts, planks)
Phase 3: Advanced AI
🧠 Personalized workout generation
🎯 Injury risk prediction
👥 Social features and challenges
📄 License
This project is licensed under the MIT License - see the LICENSE file for details.
📞 Contact & Support
📧 Email: support@mycoach.ai
🐛 Report Issues: GitHub Issues
💬 Discussions: GitHub Discussions
🏋️ Transform Your Training with MyCoachAI
The future of fitness is here - powered by AI, designed for everyone
Built with ❤️ and 🤖 by the MyCoachAI team