TryVit
A food-health app that explains its scores instead of hiding them, built for Poland and Germany.
I set the product goals, requirements, rules and acceptance checks, then decide what is ready to release. AI coding agents handle most of the implementation and technical work.

Overview
The challenge
TryVit is designed to show a score, its confidence and the factors behind it instead of only a headline grade.
Priorities
What mattered
- Product records and availability need country-aware context.
- Database access must be restricted by market and product-facing API rules.
- Health claims must stay limited while the catalog and model are still being validated.
Decisions
Key decisions
Treat market as part of the data
Country availability belongs in the product record, not only in an interface filter.
Protect product access
Row-level security and focused product APIs control how the application reaches the database.
Keep data and interface aligned
Typed contracts, validation and automated tests keep the interface aligned with the data.
Experience
What people use
Search products, inspect scores and confidence, and compare alternatives.
- Search and manual barcode lookup
- Multi-factor score and confidence
- Diet, allergen, alternatives and comparison
- Country-aware behavior for Poland and Germany
Build
How it works
Structured data and clear access rules support every score.
- Reviewed data sources and Python validation
- PostgreSQL and Supabase data rules
- Country isolation with row-level security
- Typed Next.js contracts and automated testing
Screens
A closer look
Current interface views, with source and status noted.


Current status
Where it stands
TryVit is live and actively developed. The product catalog and scoring model are still being validated.