Inspiration
Every year, millions of tons of food are wasted at the household level — often simply because people forget what’s in their fridge or pantry, or miss expiration dates.
We wanted to create a simple yet intelligent assistant that helps reduce food waste using AI and minimal manual input.
What it does
Smart Pantry Assistant is a lightweight web app that helps users track food items and their expiration dates.
It allows:
- Manual entry of food items with expiration dates
- Image recognition of common food items using the phone’s camera
- Alerts before items expire
- Recipe suggestions based on ingredients that are about to expire
How we built it
We focused on building a Minimum Viable Product (MVP) with:
- React / Next.js frontend using Tailwind CSS for clean UI
- Express.js backend
- TensorFlow.js for basic image classification (e.g., milk, eggs, fruits)
- IndexedDB or SQLite for local data storage
- Web Notifications API for local alerts
Challenges we ran into
- Adapting a pre-trained image recognition model to work in real-world kitchen conditions (lighting, angles, packaging)
- Designing a fast and user-friendly input system
- Managing offline/local data storage efficiently
Accomplishments that we're proud of
- Built a functional MVP in a short timeframe
- Successfully integrated AI for product recognition
- Designed an interface that's simple and useful for everyday use
- Addressed a real-world sustainability issue with practical tech
What we learned
- How to use Computer Vision (with TensorFlow.js) in the browser
- How to design for usability under time pressure
- Best practices in local data storage and notification handling
- How small AI-powered tools can make a big social impact
Built With
- express.js
- react
- sqlite
- tailwind
- tensorflow.js
Log in or sign up for Devpost to join the conversation.