HuskyCart

Comparing price between the scan item and other store
Scan the receipt by using OCR
Result of scanning the receipt
The pantry list of grocery
Ordering, remove grocery without login to app using Photon beta

Inspiration

Many people struggle to keep track of grocery expiration dates while managing constantly rising food prices. We aim to solve this by creating a personal AI agent that understands you, your pantry, and what matters most to your budget. By tracking groceries, predicting what you’ll need, and helping you make smarter purchasing decisions, our AI turns everyday pantry management into a more personalized, convenient, and cost-effective experience.

What it does

HuskyCart texts you suggestions from real store prices for cheaper and more nutritious options using an AI agent. You remain in control by saying whether you want to buy or ignore its suggestion, while putting the mental stress and labor in the past

How we built it

For the frontend, we used Figma to sketch out the layout. Then, we used Cursor and Claude Code to put them together and build the effects. For scanning barcodes, we used Open Food Facts, and for OCR to read receipts, we used powerful perceptron-mk1.5 model. We also used React Native and Expo to make it runs on mobile devices.

Backend we used MongoDB for our DB, Photon to send messages through iMessage, and Gemini to create suggestions based on the user's grocery goals.

Challenges we ran into

We ran into a lot of issues connecting the front end with the backend and creating an app that feels concise and fluid. It is hard to find an API that read the barcode and the receipt correctly. Also, building a database that can verify through users' email was also challenging.

Accomplishments that we're proud of

We got the OCR, iMessage agent, and database to all connect. The UI is also elegant and fit with our concept. It felt so satisfying when everything came together. Creating a nice landing page was also a huge success. We were worried our product would look like another slop vibecoded website, but we found websites with prompts for making the front end look great!

What we learned

We learned to thrive under high-stress environments and to have hope even when we are very anxious about whether our project will be completed on time. We also made connections for a lifetime through shared anxiety and how to play to each other's strengths and weaknesses to create a product that we are all proud of.