A calorie tracker for home-cooked meals — no ingredient-by-ingredient logging.
Home recipes are the worst case for manual calorie trackers: break the dish into ingredients, measure each, log them one at a time. Nommie is a calorie tracker for home-cooked meals that reads the finished plate and estimates nutrition in one shot.
Logging home cooking manually is tedious in a way that scales badly: a simple weeknight dinner can have eight ingredients, each needing its own search, weight, and entry. Nommie replaces that whole process with a photo of the finished plate. The AI identifies what's on it, estimates portion size, and logs calories and macros for the meal as a whole — no ingredient breakdown required.
What makes Nommie different
Log the plate, not the recipe
A photo of the finished meal skips the ingredient list entirely. You don't log the butter, the oil, the rice, the chicken, the sauce — you log the dish.
Handles family portions and leftovers
Photograph your own plate from a family-size dish, or log tomorrow's leftovers fresh from the fridge. Portion size comes from the plate, not a recipe yield estimate.
Works for meal prep
Cook once, photograph each container. Every prepped meal logs the same way without re-entering the recipe each time you eat it.
Learns your kitchen
Correct the AI once on a favorite recipe and it adapts to your version of it. Your bolognese is not the average bolognese — the model figures that out.
Home cooking is where traditional calorie trackers make people quit. A weeknight meal might be chicken thighs, olive oil, rice, broccoli, and a sauce with four ingredients — that's nine database searches, nine portion estimates, and nine log entries, for one plate. The cognitive cost of doing that every night is why so many tracking streaks end in the second week. A photo-based workflow cuts the entire process down to a single action.
The other advantage is accuracy. Manual ingredient logging requires either weighing every component or guessing grams on sight, and studies consistently show that human portion estimates for cooked food are 20-30% off. A vision model looking at the finished plate uses visual references for portion and returns a result that's typically within 10-15%. Counterintuitively, the faster method is also the more reliable one for home cooking.
Nommie is particularly useful for three home-cooking patterns: big-batch meal prep where the same dish is eaten across multiple days, family-style meals where individual plate portions vary, and recurring favorite recipes where the model quickly adapts to your specific version. For all three, the photo is the log — you don't rebuild the recipe each time.
Frequently asked questions
How does Nommie handle home recipes without a recipe database?+
It doesn't need one. The AI looks at the finished plate and estimates calories and macros for the dish directly, rather than reconstructing a recipe from ingredients. For home cooking, this is actually more accurate than database lookup, because your version of a recipe isn't the generic version anyway.
What about meal prep — do I log each container separately?+
Yes, and it takes seconds. Snap a photo of each prepped container and Nommie logs it individually. If you're eating the same prep for several days, the model gets faster and more accurate with each log as it learns the specific dish you're repeating.
How does it estimate portion size for family meals?+
You photograph your own plate, not the serving dish. Portion estimates come from visual references like plate size and utensils, so the model is measuring what you actually took — not the recipe yield divided by a guessed number of servings.
Can I save favorite recipes for faster logging?+
You don't need to. Because every log starts from a photo, re-logging a favorite home recipe is the same workflow as logging it the first time — and the model adapts to your specific version over time, so accuracy improves without any recipe-entry step.
Try Nommie for free
Snap a photo, get instant nutrition — no manual logging.