GoodByte | Devpost

Inspiration

Navigating the grocery store aisle has become overwhelming. Consumers today care deeply about what they put in their bodies (dietary restrictions, allergies) and the global impact of their purchases. However, reading tiny ingredient lists or researching a parent company's ethical practices on your phone while holding up the checkout line is nearly impossible. We were inspired to build GoodByte to make safe, ethical, and informed consumption of everyday products as effortless as a single tap.

What it does

GoodByte is a personalised, real-time product analysis app. Users start by configuring a profile with their specific health and dietary requirements (e.g., Nut Allergy, Gluten-Free). Using the device's native camera, users scan any product barcode and GoodByte instantly analyses the item and returns a clean, highly visual dashboard featuring a "Health Match"(Flags which display if the product is safe to consume, based on your health and dietary restrictions), a "Nutri-Score"(Grading scale that starts from A to E which symbolises the Nutritional content of the product) and an "Ethics Score"(which is calculated by our custom AI model). Rather than making users manually configure complex ethical preferences upfront, the app provides a transparent, easy-to-read summary explaining exactly how that specific item's Ethics Score was calculated (e.g., flagging deforestation risks in the supply chain or praising fair-trade sourcing). Using these visual pointers, the consumers can consume products safely and ethically by making informed decisions.

How we built it

GoodByte is a full-stack mobile application featuring a responsive cross-platform Flutter frontend and a fast, asynchronous Python/FastAPI backend. The system utilises an SQLite database managed by SQLAlchemy to store user profiles and complex dietary preferences, seamlessly routing this data to external services. By integrating the Open Food Facts API for raw barcode data and leveraging Google's Gemini 2.5 Flash alongside GNews for real-time ESG web scraping, the app dynamically processes both the health and ethical footprint of scanned products and creates a visual dashboard for the user.

Challenges we ran into

Our primary hurdles revolved around managing network latency and taming unpredictable AI outputs to ensure a smooth user experience. Because fetching data from Open Food Facts and processing complex LLM prompts takes time, we had to carefully implement asynchronous UI states and extended HTTP timeouts to prevent the app from freezing. Furthermore, we had to engineer extremely strict backend prompts and custom data validators to ensure the AI returned raw, parseable JSON rather than markdown, preventing fatal crashes in the Flutter frontend.

Accomplishments that we're proud of

We are incredibly proud of engineering the dynamic ESG Synthesis Engine, which actively resolves a product's parent company, scrapes real-time news across environmental, social, and governance pillars, and synthesises it into a standardised ethical score. Additionally, our highly contextual Dietary Engine goes beyond basic keyword matching to cross-reference ingredients against a user's unique combination of allergies, dislikes, and dietary goals. We matched this technical complexity with an elegant UI, creating engaging loading dialogs that keep users entertained during high-latency AI tasks.

What we learned

Building GoodByte reinforced the critical importance of defensive UI programming and taught us that prompt engineering is a vital backend discipline. We learned how to carefully manage asynchronous navigation states in Flutter, using mounted checks to prevent memory leaks or crashes when users navigate away during an active API call. Ultimately, we discovered that strictly defining LLM constraints and output schemas is just as essential to system stability as writing robust database relationships and managing cross-origin local networking.

What's next for GoodByte

Looking ahead, the next phase for GoodByte focuses on scaling our impact and deepening user engagement. Our primary goal is to introduce a "Smart Alternatives" engine that doesn't just flag problematic ingredients or poor ethical scores, but actively recommends healthier, more sustainable substitute products. On the technical front, we plan to optimize our dynamic ESG tracking by implementing a robust caching layer to reduce AI latency, while migrating our backend from a local SQLite setup to a fully scalable cloud database to support a rapidly growing user base. Ultimately, we envision expanding GoodByte into a community-driven platform featuring gamification, where users can earn badges, track their long-term footprint, and celebrate their positive impact on both personal health and the planet.