🚀 Inspiration

Food wastage is a silent crisis, especially at the household level. While many apps track expiry of labeled items, no system exists to predict expiry for unpacked items like vegetables and fruits. After researching, we found that most existing apps rely on hardcoded data and lack proactive user alerts. Inspired to solve this, we built a solution that predicts expiry using real-time inputs, notifies users before spoilage, and reduces waste sustainably.


💡 What it does

SmartCity FoodExpiry AI helps users track their kitchen inventory smartly. Users can:

  • Add items via manual input, barcode scan, or voice
  • Get AI-powered expiry predictions for unpacked foods
  • Receive SMS alerts before items expire
  • Discover recipe ideas or donation/composting tips
  • Get practical suggestions to extend the shelf life of food items
  • Analyze inventory health via visual reports
  • Access a restaurant dashboard for bulk inventory tracking (under development)
    -Get AI-powered expiry predictions for unpacked foods by entering product details such as item type, storage conditions (like fridge, freezer, or room temperature), and item condition (fresh, opened, etc.), then clicking the Predict button to instantly receive an estimated expiry date ---

🏗️ How we built it

  • Frontend: Built with React.js and TailwindCSS for a clean, responsive UI
  • Backend: Node.js with Express and MongoDB to manage data, authentication, and Twilio SMS
  • ML Expiry Prediction: XGBoost model trained on a custom dataset generated via Mimesis
  • Voice Integration: Web Speech API (Speech-to-Text) enables hands-free input to reduce manual errors
  • Visualization: Python-based data visualizations show inventory health and patterns

🧗‍♂️ Challenges we ran into

  • No existing Indian dataset for food expiry, especially unpacked items
  • Hard to generalize expiry dates for perishable items with varying storage/condition
  • Handling different user inputs and minimizing errors in expiry prediction
  • Designing a UI that balances usability with functionality

🏆 Accomplishments that we're proud of

  • Developed a working AI-based expiry prediction system for non-labeled items
  • Built a scalable inventory system with SMS alerts and health reports
  • Integrated real-time voice input for ease and accessibility
  • Created a custom dataset using mimesis after finding no suitable open datasets

📚 What we learned

  • Importance of data quality and real-world relevance in machine learning
  • How a small feature like voice input can greatly improve UX
  • That solving real problems requires not just coding, but domain research and user empathy

🔮 What's next for SmartCity FoodExpiry AI

  • Expand dataset with real-world records like harvesting dates, weather, and location
  • Train on regional produce types for more accurate predictions
  • Add multilingual support for voice and interface
  • Build features like family inventory sharing, local donation suggestions, AI-powered recipe generation, and personalized tips to extend food shelf life
  • Complete and launch the restaurant dashboard for bulk users and better inventory management

Built With

Share this project:

Updates