Estimates, not medical measurements
Calories, macros, and micronutrients are presented for everyday tracking. They should not be used as medical diagnosis or treatment guidance.
Back to NutriOn
Food data and accuracy
Nutrition data is most useful when its limits are clear.
NutriOn treats AI food recognition as a starting point. The app is designed to help users review estimates, correct context, and track patterns over time.
Accuracy principles
A safer way to explain AI nutrition tracking
NutriOn avoids presenting AI output as a perfect measurement. The product experience should make uncertainty understandable and correction easy.
Visible review before saving
Users can correct portions, food names, and preparation details before the result becomes a saved record.
Trends matter more than one perfect meal
Daily and weekly patterns are usually more useful than treating a single estimated meal as exact.
Limits are stated plainly
Mixed dishes, hidden ingredients, lighting, and serving size can all affect AI-assisted estimates.
When to review extra carefully
Food photos can miss details that matter. These cases should be treated as prompts to review the estimate, not as failures of the user.
- Mixed dishes with hidden oils, sauces, or toppings.
- Restaurant meals where serving sizes are unknown.
- Low light, cropped photos, or overlapping food items.
- Medical diets, allergies, or disease-specific nutrition needs.
NutriOn
Track meals with context, not blind trust.
Open the NutriOn download section and keep the same attribution context for install and subscription measurement.