Case study · iOS and Android
BioBrief: Food Scanner
A mobile product that makes dense food labels easy to understand at the moment a shopper has to decide, now in six languages and available in the United States.

- Product
- CodingShadows product
- Scope
- Mobile product, AI analysis, Go backend, product UX
- Platform
- iOS and Android
- Outcome
- 4.8★ · 6,000+ store ratings
01
Problem
Food packaging compresses ingredient lists and nutrition data into a format that is hard to read quickly, and barcode databases miss many products, especially outside Romania. The product had to go from a photo or a scan to a useful explanation without overwhelming the user or presenting the result as medical advice.
02
Scope and constraints
- • A fast camera flow that reads the ingredient list itself, with the barcode as a shortcut
- • Clear separation between product facts and generated explanation
- • Useful states when a product is incomplete or cannot be recognized
- • One product in six languages, Romanian first, and a US market where photo scanning is the only path
03
Implementation
The experience centers on one action: photograph the ingredient list or scan a barcode, then read a health score with additive and allergen warnings. History, ingredient details and a shopping list stay one tap away without competing with that path.
Geneva, the in-app AI nutritionist, answers follow-up questions by text or by voice, AI Chef turns products into recipes, and a calorie calculator rounds out the daily tools.
In August 2026 the app opened to the United States with photo-only scanning, and it now runs in six languages: Romanian, English, German, Hungarian, Italian and Polish. The backend moved from Spring Boot to Go on PostgreSQL, product search included, and biobrief.app became the English home of the product.
- Flutter
- Go
- PostgreSQL
- On-device text recognition
- LLM analysis
- Six locales
04
Results so far
4.8★ · 6,000+ store ratings
More than 300,000 downloads and over 6,000 store ratings: 4.8 on Google Play from 4,210 reviews and 4.92 on the Romanian App Store from 2,417. The product has been covered by Euronews Romania and by iLikeIT on PRO TV, independent evidence that it reached well beyond an internal prototype.
Evidence: Google Play and Apple App Store Romania rating data checked 3 October 2026; Euronews image from the CodingShadows media archive.