LogmaYou log everything for a day and the app says 2,000 calories. Then you add it up yourself from the labels and land closer to 2,850. That exact story shows up in an App Store review, and it is how a lot of people find out their log was off all along.
Usually the app did the math right. The numbers it was given were the problem: an entry someone typed in wrong, a cooked food logged with raw values, a serving that meant something else to whoever created it.
Quick disclosure: I'm Batu, and I build Logma, a calorie tracker for iPhone you talk to. So I'm biased. Everything about other apps below comes from their own App Store pages and their newest low reviews.
“A cup of white rice, 150 grams of grilled chicken breast and a side salad with a tablespoon of olive oil.”
Example: Logma splits this into rice, chicken breast, salad and olive oil and logs about 600 kcal with macros.
How often people complain about wrong food data
In November 2026 I read the newest 1 and 2 star US App Store reviews of 12 calorie apps, 2,066 reviews in total. About 9% of them mention wrong food data or bad database entries. Here is that share for six of the apps.
| App | 1 and 2 star reviews read | Mention wrong food data |
|---|---|---|
| Lifesum | 282 | about 18% |
| MacroFactor | 129 | about 16% |
| Cal AI | 241 | about 12% |
| MyFitnessPal | 209 | about 7% |
| Yazio | 284 | about 6% |
| Lose It | 141 | about 3% |
These are only the unhappy reviews, and every app here is rated 4.6 stars or higher on the US App Store. A high share does not mean the worst database, only that unhappy users of that app bring up wrong numbers more often.
The complaints sound alike from app to app. One reviewer scanned a protein shake whose label said 30 g of protein and saw 106 g logged. A reviewer in Türkiye searched for turkey breast and found one entry at 104 calories per 100 g and another at 134. As one App Store review put it, the database was "full of wildly inaccurate entries."
Why the same food has five different numbers
Big food databases grow fast because users can add to them, which is also why many entries were never checked by anyone.
- Anyone can add an entry. In apps that accept user entries, a typo in the calories field or a missing fat number goes live as soon as it's saved.
- Copies of copies. People copy an existing entry to make their own. If the original was wrong, every copy is wrong too, and more copies make a wrong pick more likely.
- Raw vs cooked. 100 g of dry white rice is about 360 calories, and 100 g of cooked rice about 130, because cooked rice is mostly water. Chicken breast goes the other way: about 120 calories per 100 g raw and about 165 cooked. Weigh one and log the other and you're off by a lot.
- Per 100 g vs per serving. One entry says "1 serving" and means 100 g. Another means one cup, a third the whole package. In the search results they look the same.
- Brand vs homemade. A restaurant lasagna, a frozen one and your own can be hundreds of calories apart per plate, under the same name.
- Barcodes point to typed entries. In many apps, a barcode leads to an entry someone typed in from the package. If they typed it wrong, picked the wrong serving, or the recipe changed since, the scan returns wrong numbers that look official.
How a few wrong entries add up to 850 calories
Small errors stack. Here is a normal day where every log looks reasonable and the total still comes out far too low:
- Rice: an entry where one serving means 100 g, for a bowl of about 250 g cooked. About 195 calories missing.
- Chicken breast: 150 g weighed after cooking, logged with a raw entry. About 70 missing.
- Olive oil: two tablespoons in the pan, never logged. About 240 missing.
- Peanut butter: logged as one tablespoon, but you had two. About 95 missing.
- Dinner out: a pasta entry from someone's home recipe instead of a restaurant plate. About 250 missing.
That's roughly 850 calories on a day you logged everything, the same gap as the reviewer above, and enough to wipe out a planned calorie deficit. The fix is not logging more. It's making the entries you use often trustworthy once.
How to protect your numbers in any app
None of this needs a new app.
- Pick verified entries when the app marks them. Some apps flag checked entries, and some describe their whole database that way: Cronometer's App Store description says it is built on a curated database of more than 1.1 million verified foods, and MacroFactor's calls its food database verified.
- Check new entries against the label once. The first time you log a packaged food, compare calories, protein, carbs and fat with the package. If they match, save it as a favorite.
- Skip entries that look lazy. Round numbers, missing macros, or calories that don't fit the macros are signs someone guessed. Protein and carbs are about 4 calories per gram, fat about 9.
- Weigh once to calibrate. Weigh your usual bowl of rice, pour of cereal and scoop of peanut butter one time each. After that you know what "my bowl" means in grams. More in tracking calories without weighing food.
- Match raw with raw and cooked with cooked. If you weigh it cooked, pick an entry that says cooked.
- Log cooking oil as its own item. A tablespoon is about 120 calories, and it's the item most often missing from a log.
- Watch the trend, not one day. If your weight over three or four weeks doesn't match your log, your entries or your target are off. The calorie calculator gives you a fresh starting number.
How Logma skips the crowd database
When I built Logma, I left out the part where you scroll user entries and guess which one is right. There is no user made food database to pick from. You say or type what you ate in one normal sentence, and Logma splits it into items and estimates each one from what you said: the food, the amount, how it was cooked.
“About two cups of homemade beef chili with a handful of shredded cheddar and a spoonful of sour cream.”
That becomes three items with calories and macros in about 3 seconds, without twenty entries called chili to choose from. If you cooked with oil, say so and it gets counted. If a portion is off, you correct it in words, for example that it was closer to one cup.
To be clear, this is an estimate, like every calorie number in every app. With a crowd database, the risk is picking someone else's mistake. With Logma, it's how well you describe the amount, so say grams when you know them.
Where it fits less: if most of what you eat comes with a label, an app with a verified barcode database is the better tool, since Logma doesn't have a barcode scanner. For the whole field side by side, see my calorie tracking apps comparison for 2026.
Logma is iPhone only for now, with Android coming on November 14, 2026, and it has no ads. The yearly plan starts with a 3 day free trial, then it's $49.99 a year, or $9.99 a month.
FAQ
Why does my calorie app show different calories for the same food?
In many apps the entries come from users, so the same food is added many times: raw or cooked, per 100 g or per serving, branded or homemade, and some entries copy a mistake. Pick one marked verified, or one that matches your label.
Why is the nutrition wrong when I scan a barcode?
A barcode only points to an entry. If that entry was typed in wrong, set to the wrong serving, or made before the recipe changed, the scan returns wrong numbers. Compare the first scan with the label, then fix or replace the entry once.
Should I log food raw or cooked?
Either, as long as the entry matches how you weighed it. Dry white rice has about 360 calories per 100 g and cooked rice about 130, so mixing them up is one of the biggest single errors you can make.
Which calorie app has the most accurate food database?
Reviews can't give you a fair ranking. Cronometer and MacroFactor both describe their databases as verified in their App Store descriptions, which is a good start. In any app, portions and forgotten oil usually cost more than the database.
Does Logma use a food database?
Not one you scroll. There is no user made database to pick from. Logma estimates each item from what you said, including the oil or sauce you mention, and you can correct a portion in words.
Sources: Cronometer on the App Store · MacroFactor on the App Store · Lifesum on the App Store · Cal AI on the App Store · Logma on the App Store. Numbers are approximate and can change; nothing here is medical advice.
