General information, not medical advice — your care team knows your labs and your limits. Full disclaimer.
Most apps in this space are trackers. You cook something, then you tell the app what you cooked, and it tells you how you did. That works if your only problem is awareness. It works badly if your problem is the question at 5 p.m. — what do I cook tonight that fits a potassium limit, a protein floor, and three other people at the table?
A planner answers that question before you shop. The distinction matters most for people who have a condition to respect, because a wrong dinner is not just a bad log entry. Here is what to check before you trust any app with that job — including this one.
Open the app and look for the decision. Can it hand you a week of dinners without you naming a single dish first? If the main screen is a diary, a barcode scanner or a photo button, it is a tracker with a recipe section — useful, but it will never take the 5 p.m. question off your plate.
The five-minute test: install it, answer the setup questions, and see whether you finish holding a shopping list you could actually take to a store.
Ask an app where a number came from and you learn everything. For most condition diets the hard numbers are not calories — they are sodium, potassium and phosphorus, and those are exactly the ones AI models state most confidently and least accurately.
What good looks like: the app can show you the source food behind a number, not just the number. Plate Planner matches each ingredient to an entry in a bundled USDA-derived food dataset and will show you that breakdown on request — including the lines it could not match, which it says plainly rather than quietly estimating.
The five-minute test: pick any nutrition figure and try to trace it. If nothing explains it, treat every figure in that app as a sketch.
A calorie target is not a kidney diet. Renal planning needs potassium and phosphorus per serving, not only sodium. GLP-1 planning needs a protein floor to help protect muscle while you eat far less, and gentler days around an injection. Diabetes planning cares about the shape of the carbohydrate, not only the amount.
The five-minute test: turn the condition on and see whether the food itself changes — different dishes, different limits, a visible ceiling — or whether the app simply prints a badge and carries on.
This is the failure nobody markets: the plan fits you and nobody else, so you cook twice. It is one of the most common reasons a condition-aware plan is abandoned in month two.
What good looks like: one plan that respects the strictest need at the table and scales portions for everyone else — a kidney-friendly plate and full portions for the rest of the family, out of the same pan and the same grocery list.
Recipe repetition is the most universal complaint in this whole category. Plenty of planners are delightful for eight days and stale by day twenty, because nothing stops them serving the same six dinners forever.
What good looks like: an app that remembers everything it has ever served you and refuses to repeat it. Plate Planner keeps a permanent record of every dish it has planned for you and will not serve the same dinner twice — a claim verified by generating four consecutive weeks, more than a hundred dinners, and checking every one for repeats.
The five-minute test: generate two or three weeks back to back and read the dinner list.
Condition data is sensitive by definition. Look for plain answers: what leaves your device, whether you can export and delete everything, whether cancelling is a button or an email, and whether the free tier is genuinely usable or a demo with a wall.
Plate Planner plans a whole week without an account, keeps your plan on your device unless you sign in to sync it, shows no ads, sells nothing about you, and has export and delete built in.
They solve different problems. A tracker tells you what already happened; a planner decides what happens next. If you cook most of your meals and have a condition to respect, planning removes the decision that causes the trouble — and logging mostly stops being necessary, because a plan you cooked is a plan whose nutrition is already known.
Only as far as its numbers can be traced. A language model will state a potassium figure with total confidence and no source. The safer version grounds each ingredient in a real food database and tells you when it could not. Ask an app to show its working before you rely on it, and keep your care team in the loop regardless.
Whether it produces a shopping list without you naming the dishes. Everything else — traceable numbers, condition depth, household portions, variety, privacy — matters, but an app that cannot make the decision for you is not a planner.