NutriSense

NutriSense — Snap your fridge. Shop smarter. Eat better.

Comment

💡 Inspiration

Many people struggle with:

  • Food waste from forgotten fridge items
  • Grocery shopping without nutrition awareness
  • Cooking inspiration despite having ingredients

NutriSense was inspired by the idea that your fridge already contains valuable data—AI just needs to understand it.

🚀 What It Does

NutriSense guides users through a seamless workflow:

  1. Scan your fridge using a photo
  2. Detect ingredients with AI, including confidence and freshness
  3. Understand your health goals and diet preferences
  4. Recommend what to buy, what to cook, and where to shop nearby
  5. Continuously adapt as your health data and goals change

✨ Features

🧠 Intelligent Ingredient Detection

  • Upload or drag-and-drop fridge photos
  • AI-powered ingredient recognition with confidence scores
  • Freshness tracking: Fresh, Use Soon, Expired
  • Manual confirmation, editing, and deletion
  • Visual confidence indicators

👤 Personalized User Profiles

  • Track age, gender, height, and weight
  • Health goals:
    • Maintain Weight
    • Lose Weight
    • Gain Muscle
    • Increase Energy
  • Diet preferences:
    • Balanced, Vegetarian, Vegan, Keto, Paleo, Mediterranean
  • Custom taste preference tags (e.g. spicy, Italian)

🛒 Smart Shopping Recommendations

  • Nutrition-aware grocery suggestions
  • Ingredient gap detection
  • Nearby store listings with distance
  • Price comparison across retailers
  • Shopping lists grouped by store

🍽️ Recipe Suggestions

  • Recipes based on available ingredients
  • Cuisine and taste filtering
  • Recipe metadata:
    • Cook time
    • Difficulty
    • Calories
  • Visual recipe cards

🛠️ How We Built It

NutriSense is built as a modern, lightweight web application focused on usability and clarity.

Technology Stack

  • HTML5 — Semantic markup
  • CSS3 — Responsive design, gradients, and animations
  • Vanilla JavaScript — Frontend application logic
  • Supabase — Authentication, PostgreSQL database, and storage
  • FastAPI — Backend API and business logic
  • Gemini — AI-powered vision and language understanding
  • Presage — Predictive intelligence and recommendation logic

Challenges We Ran Into

Building NutriSense within a hackathon timeframe came with several challenges. One major hurdle was designing an AI-ready architecture without an actual backend, while still keeping the product realistic and scalable. We also had to balance rich personalization (multiple health goals, diets, and preferences) with a clean, intuitive user experience that wouldn’t overwhelm users.

Another challenge was making AI outputs trustworthy and explainable. Ingredient detection needed confidence scores and manual correction flows so users could understand and control what the AI was doing. Finally, simulating real-world data—such as store pricing, locations, and recipe metadata—required careful design to feel authentic without external APIs.

Accomplishments That We're Proud Of

We successfully built an end-to-end kitchen intelligence experience—from fridge scanning to shopping and recipe recommendations—entirely in the frontend. The app demonstrates a realistic AI workflow with confidence scoring, freshness tracking, and explainable recommendations.

We’re especially proud of:

  • A polished, responsive UI that feels production-ready
  • A clear, scalable data model designed for future AI and backend integration
  • Thoughtful UX decisions that prioritize transparency and user trust
  • A complete demo that communicates the product vision effectively

What We Learned

Through building NutriSense, we learned how to design AI-driven products beyond just the model itself. Data structure, explainability, and user feedback loops are just as important as accuracy.

We also gained hands-on experience in:

  • Modeling time-based health and nutrition data
  • Designing AI-friendly user workflows
  • Rapidly prototyping complex ideas under time constraints
  • Making tradeoffs between technical depth and user experience

What's Next for NutriSense

NutriSense has strong potential to grow into a fully intelligent kitchen assistant. Next steps include:

  • Launch a native iOS and Android app with camera-first scanning, push notifications, and offline-friendly fridge tracking.
  • Integrate with local online grocery stores to automatically place orders, schedule deliveries, and optimize purchases using AI-driven pricing, nutrition, and availability insights.
  • Implementing nutrition analysis (macros and micronutrients)
  • Tracking health progress over time with adaptive recommendations
  • Adding smart alerts for expiring food and nearby price drops
  • Learning from user feedback to continuously improve suggestions

Our long-term vision is to make NutriSense a trusted daily companion for healthier, smarter eating.

Built With

Share this project:

Updates