Inspiration

This hackathon’s theme, “Let’s Dig In,” is also a casual way of saying “Let’s eat.” That got us thinking about how little most of us know about what we’re actually eating, and how it affects us. A nutrition label lists the numbers, but it doesn’t clearly tell you whether a food fits your goals for the day. An ingredient list names the allergens, but it doesn’t say whether your child can safely eat it. When there’s no label at all, like a home-cooked meal or a restaurant plate, most people are left guessing.

We wanted to solve that by turning what’s on your plate into a clear and honest answer: how many calories and macronutrients you’re eating, whether something conflicts with your allergies or diet, and whether it’s safe for the people you cook for. So we dug in, and HonestPlates was born.

What it does

HonestPlates is a free iOS mobile application that's built for everyone who wants to know what they are eating. The app includes features for tracking their calories and macronutrients, barcode, nutrient label, and plate scanning, an allergy and dietary flag, and detailed flags for ingredients found within the food. This app is built for the convenience of food information and logic for people who want to hit protein goals, limit harmful ingredients, and quickly check if the food on the plate meets their dietary needs.

How we built it

HonestPlates is an iOS app built with React Native and TypeScript in Xcode. We split the work into three parts: reading food, judging it, and tracking it.

– Reading food: The phone's camera feeds Google ML Kit's device text recognition. The ML kit can either choose to parse the text or reads the barcode, allowing the app to look the product up in the Open Food Facts database. Photos of a meal got to an AI vision model through a small server we run, to estimate its portions and nutrients.

– Judging it: An ingredient parser pulls the ingredients list out of the scanned detects and goes through a personal allergen algorithm which checks each user's dietary restrictions. The same checks run for every household member so one scan shows who at the table can eat it. The app also pulls from openFDA API once a day to match them by barcode against foods the user has logged, so they’re alerted if something they ate gets recalled later.

– Tracking it: Entries are saved only on the phone (Zustand state saved to AsyncStorage). There are no accounts, and health data never leaves the device. The daily calorie ring and weekly charts are drawn with react-native-svg.

Challenges we ran into

When working on this project, one challenge was trying to find something that would differentiate us from the rest of the market. These are many apps on the App Store that have similar features to HonestPlates, but we found that no other app centralized each of these features to create a full health and nutrition solution.

Accomplishments that we're proud of

Working on HonestPlates, we're proud of our feature to be inclusive for all people with our robust algorithm which can accurately check if the food ingredients match all kinds of dietary needs. We’re also proud that one scan answers a question no regular calorie tracker does: whether a food is safe for everyone at the table, from a parent counting calories to a child with a milk allergy. And we’re proud that it does all of this while keeping every user’s health data on their own phone.

What we learned

With this project, we learned how to turn real-world data into reliable information, and how to structure and build mobile apps. Public databases like the Open Food Facts and FDA's recall data are powerful but have limited power in a vacuum–something our app helps with by interpreting that data meaningfully for its users. We also learned the practical side of building for iOS with React Native by implementing camera and native libraries, working with Xcode signing and builds, and coordinating a team through GitHub branches and pull requests to work effectively without interfering with each other's work.

What's next for HonestPlates

With HonestPlates, the next step would be further refining our current features and user experience. That would include improving our current AI server model to better identify food and adjusting the food safety grading system to be more strict.

Share this project:

Updates

Submission history