Know what you have. Love what you cook.

Inspiration

If you're a time-crunched student who still wants meal variety, you want to spend time cooking recipes instead of spending time searching for them. PantryPal takes the overload of recipes available online and tailors recipes to you. It takes your taste preferences and stock of items you already have in your fridge or pantry so you can get started faster, reduce food waste, and explore your culinary abilities. Make cooking fun again, and let PantryPal keep track of your ingredients, food preferences, and recipe search so you can focus on the meal.

What it does

An app that tracks the ingredients in your household and generates recipes from that set and additional ingredients to purchase. It would track ingredients from receipt images and track usage via the recipes suggested and used to have an up-to-date inventory.

Features

This app is designed for beginner or budget-minded cooks who are overwhelmed by the information available online. The app recommends recipes that utilize what's already in the pantry, and it keeps track of remainder ingredients that need to be purchased. Users can save money by fully utilizing their inventory (e.g. no more throwing out that barely-used can of tomato paste or that sad bag of spinach) and will eat healthier with the comfort of a home-cooked meal.

The suggested recipes can be tailored to the user both via user input and app learnings about the user. Users can input what type of cuisine they want, calories or macros they’re targeting, dietary restrictions, budget restrictions, and occasions for cooking (e.g. date night). The learnings come from understanding the difficulty and techniques accomplished in completed recipes, tolerance and preferences (e.g. too spicy / salty), and age of inventory (e.g. could generate recipes for nearest expiring ingredients).

How we built it

After settling on this idea after a brainstorming session, we white-boarded the user journey and the key pages of our product. After solidifying our desired layouts, we used Gemini to hone our prompts for Lovable to generate the front-end site. Once we had a draft for the app, we user-tested it to find areas of improvement, writing down necessary changes in a wishlist format and running Lovable and Gemini as-needed. Additionally, we needed to optimize the initial output for mobile and tested it by viewing the app on a mobile browser.

Challenges we ran into

Using Lovable's built-in AI is credit-hungry, forcing us to think about using it efficiently to conserve credits. We needed to top up the balance with $10 worth of credits.
We've had issues with preference persistence, as a checkbox would not remain checked if we switched between webpages. This is likely due to insufficient integration between the database and front end which we could reconcile with more time.

Accomplishments that we're proud of

Planning out the user journey and UX so that we were efficient in implementation and troubleshooting in Lovable
Successfully identifying what to iterate on in the design process

What we learned

Troubleshooting code with an LLM is convenient but the results were uncertain
How to use Lovable

What's next for Pantry Pal

Integrating an LLM to generate recipes
Integrating an OCR (Optical Character Recognition) to ingest the receipt images and converting line-items into items for the inventory database
Building 3rd-party API integrations to enable one-click ordering on Instacart, Walmart+, Amazon Grocery, etc. to make grocery shopping even easier