A photo handles most of it. When a dish hides its ingredients, describe it or scan the barcode instead.
Point your camera at the plate. The AI separates each component and returns calories and macros.
Type what you ate — “two eggs and toast.” Best for soups, sauces, and anything the camera can’t see inside.
Point at any packaged product for an exact match from a global food database.
Your camera frames the meal and the depth sensor estimates portion volume.
A multimodal model trained on thousands of food images separates each ingredient and its proportion.
Calories, protein, carbs, and fat land in your daily log with insights and reminders to keep you on pace.

Photo recognition lands around 80% accuracy on everyday meals. If a result looks off, tap Correct results — it fixes your entry and teaches the model to do better next time.