Food tracking has a famous retention problem. People start, log diligently for a fortnight, and stop. The usual explanation is discipline. We think the tools are built wrong.
Two assumptions that do not survive contact
- That food comes with a barcode. A plate of home-cooked food has no packet. In most of the world most meals are cooked at home, so the barcode database is empty exactly where it is needed.
- That people will weigh things. They will not, past the first week. Any system requiring grams is a system with a two-week lifespan.
What we do instead
One photograph. It identifies what is on the plate, including dishes that only exist in particular kitchens, and shows the result in proportions rather than in weights — how much of the plate is what, rather than a number that implies a precision a photograph cannot deliver.
Honest about precision
A photograph cannot tell you how much oil went into a curry. We do not pretend otherwise. What it gives you is a usable picture of what you are actually eating, which for almost everyone is far more than they had before, and which does not collapse the moment you eat something that is not in a database.
What it will not do
It will not tell you what to eat because of a value on your blood test. Connecting a laboratory result to a dietary instruction for a specific person is a clinical recommendation, and it is not ours to make.
This article explains how our technology works. It is health information, not medical advice.