NutriWise

Overview

Barcode-scan food analysis with a Gemini verdict tailored to the user's health context and daily intake.

Problem

Nutrition databases often lack complete ingredient data and give generic answers, not guidance tailored to the person and their day.

What I built

• Barcode scanning of packaged foods.
• Gemini fills ingredient gaps when the food database API lacks details.
• A personalized verdict covering rating, additive hazards, and appropriate portion size, plus authentication and user flow.

Architecture

Mobile client → food database API lookup by barcode → Gemini 2.5 Flash prompt combining user context, scanned item, and daily intake → structured verdict shown in the UI.

Key technical decisions

Challenges and how I solved them

• Missing ingredient data from the food database API, handled with model-based gap filling.

Outcome

Working MVP with personalized food verdicts; APK built.

What I would improve next