Why estimates vary
Two plates can look identical and differ by hundreds of calories. Cooking oil, butter, creamy sauces, dressings, and drinks are hard to see in a photo. Portion depth is hard to judge from a single angle. Packaging and restaurant recipes vary. Chompino’s AI returns a plausible range based on what it can see or parse — not a weighed inventory of every gram.
When the estimate is usually good enough
Clear photos of familiar home meals, simple plates, and common foods tend to work best for awareness and daily logging. If you mainly want to stay consistent — not hit a clinical target to the calorie — a solid ballpark plus your own tweak is often enough. That is the product’s design: fast capture, human confirm.
When you should override
Override when you know the meal was oilier, larger, or denser than it looks. Override for drinks, desserts, and shared plates. Prefer manual entry when you have a package label or a known restaurant entry. Use description mode when the photo is dark, cropped, or messy. The confirm step exists so the AI never has the last word.
What Chompino is not
Chompino is not a medical device and does not diagnose or treat conditions. It is not a substitute for a registered dietitian if you need clinical nutrition care. We would rather say that clearly than promise perfect accuracy. For the product flow, see how it works; for what you can do in the app, see features.