Calorie counting assumes a normal appetite. On a GLP-1 medication such as Wegovy, Zepbound or Mounjaro, the daily question changes. Many people eat far less than they used to, so the problem is rarely eating too much. It is whether the small amount they do eat is doing enough. A joint advisory from four nutrition and obesity organizations lists adequate protein as a priority during GLP-1 therapy, and a calorie total cannot tell you whether a 300-calorie lunch delivered any. That gap is why a new kind of feature is appearing in tracking apps: a score for each meal.
What a meal score does
Instead of one running total, a meal score reads the food you logged and judges it on the things that matter when appetite is small: protein relative to calories, fiber, how processed the food is, and the portion. The result is a number from 0 to 100 and a letter grade.
The useful design detail is that the grade describes the food, not the serving. Nutrients are normalized to a fixed basis before any threshold is applied, so a bigger plate of the same meal gets the same grade with more calories. And the thresholds come from published nutrient criteria rather than a formula fitted to look right. It is an educational signal, not a medical one.
What the numbers look like
These examples are published by GLP AI and were generated by the app’s own scorer, not written by hand.
- Plain nonfat Greek yogurt, 3/4 cup: 87. Add half a cup of strawberries and a tablespoon of chia seeds: 100. Protein and fiber in the same bowl.
- Grilled chicken breast, 4 oz: 87. Two pieces of fried chicken: 57. The protein is still there, but the breading, saturated fat and sodium pull it down.
- Oatmeal cooked with water, 1 cup: 85. A packet of instant maple and brown sugar oatmeal: 60. Plain oats with a scoop of whey and blueberries: 98.
- Granola, half a cup: 25. The breakfast that looks healthy carries a lot of sugar and fat for the fiber it offers.
- A medium banana: 78. Two scrambled eggs: 70. A cup of sliced strawberries: 89.
The pattern is simple. Protein and fiber raise the score; sugar, sodium and heavy processing lower it. Those are the same priorities the advisory describes, applied one plate at a time.
How it works in practice
You log a meal by photo, barcode or a typed description. The app identifies the components, estimates the portion, and the score lands on the saved entry. If the estimate is off, you correct the serving or an ingredient and the score recalculates. GLP AI, for example, scores every logged meal this way and keeps meals, doses and side effects on one timeline, so a low-appetite stretch after a shot shows up next to what you managed to eat.
What a score cannot tell you
A meal score does not know your dose, your side effects or your personal targets. Portion estimates from a photo are approximate until you confirm them. It cannot say whether you ate enough across a day, only how well each meal was built. Agree your protein and calorie targets with your prescriber or a dietitian, and on a day when nothing sounds good, a small meal with a middling score still beats skipping.
In short
Calories measure quantity. On a GLP-1, quality per bite is the harder problem, and a per-meal score is one practical way to see it.
