What Platey is, and what it refuses to be

Platey is an AI food diary and macro tracker for people who train. You photograph a meal or describe it in a few words, and it estimates the calories, protein, carbohydrate, fat and fibre, tracks your day against your targets, and adjusts around your training — running, cycling, weights, CrossFit or football.

It is a British product. Serving sizes, foods and language are all UK — a pint, a meal deal, a Friday takeaway, parkrun on Saturday morning. It runs in a browser, sits on your home screen like an app, and there is nothing to install.

Almost everyone who has tried to count calories has abandoned it, and almost always at the same point: the search box. Finding the right entry for “chicken curry, homemade” at nine at night, then choosing between fourteen near-identical results and a dropdown of serving sizes, is a chore nobody sticks at for long.

So the search box is gone. You photograph it or you say it, and something else does the fiddly part. That one decision is why the rest of the app looks the way it does — no barcodes, no gram scale, no database to maintain, and an honest admission that the numbers are estimates.

The second decision was to make the arithmetic boring. The AI identifies food; it never adds anything up. Every figure you see is worked out by plain, tested code, which is what makes the totals trustworthy even when the estimate underneath is approximate.

Nothing turns red. Nothing says failed. No food is described as bad, clean, naughty or a treat, and there is no language anywhere for missing a run. Targets are described as a guide, and being over one is information rather than a verdict.

This is a design rule rather than a preference, and it is enforced by tests: hundreds of lines of the app’s copy are generated and checked automatically for judgemental language, gym-bro phrasing and sports-science jargon before anything ships.

No barcode scanner. No food database to search. No streaks, badges or leaderboards. No training plan, no recovery score, no fitness-and-freshness chart. No meal plans, no shopping lists, no supplement recommendations. No ads, and nothing sold about you.

Some of that is a roadmap question and some of it never will be. The training side in particular exists to make the food advice better, not to grow into a second app.

A photo estimate of a home-cooked meal will not be right to the gram, and it is worse with mixed dishes than with a plate of recognisable things. If you need accuracy for a medical reason, weigh your food — Platey is the wrong tool and would frustrate you.

It is also new. There is no long track record to point at, no user count worth quoting and no reviews, and it would be easy to invent all three. Judge it on one logged dinner instead — that takes a minute and tells you more than a testimonial would.