NutrAI

Inspiration

Health-conscious individuals need personalized nutrition advice. AI and personal health data can provide tailored meal plans.

What it does

• Generates personalized meal plans based on user data
• Tracks calorie intake using computer vision
• Provides ongoing recommendations based on progress

How we built it

• Frontend: React
• Backend: Node.js, and a local LLM server
• AI Models: Qwen 0.5B LLM, YOLOV8, Florence-2
• Local data storage for privacy

Challenges we ran into

• Integrating multiple AI models for mobile
• Balancing comprehensive data collection with user experience
• Ensuring accurate food recognition and nutritional information
• Generating balanced, personalized meal plans

Accomplishments that we're proud of

• Implemented local AI models on mobile devices
• Created a user-friendly interface for data input and meal planning
• Developed vision-based meal identification
• Designed an adaptive meal planning system

What we learned

• Implementing privacy-preserving features
• Optimizing AI models for mobile
• Translating nutritional science into personalized advice
• Creating personalized user experiences

What's next for NutrAI

• Expand recipe and ingredient database
• Implement advanced health data integration
• Develop social features
• Partner with nutrition professionals
• Improve AI models for better accuracy

Built With

• javascript
• local-llm
• node.js
• python
• react
• tailwind