Calorie tracker for iPhone: know what you eat. Including what nobody wrote down.

Nyamly scans a barcode and reads back the published nutrition data. Where a figure was never published, it says so, instead of writing a zero.

Today screen: 950 kcal left of a 2,000 kcal limit, meals by breakfast, lunch and dinner, macro bars beneath.

Scan a barcode. Read the report.

Point the camera at the barcode. Nyamly reads back the label and lays out three things: the Nutri-Score, the NOVA processing group, and the UK nutrient-profile score.

This candy bar publishes every input the UK model needs — sugar, saturated fat and sodium all at the worst cut-off, no fibre or fruit content to offset them. So the score reads 0, rated Bad: an exact result, not an estimate.

Nutri-Score E. Ultra-processed. Every input is on the label, so this is an exact score, not a floor.

“No score” is an answer.

This trail mix bar has no published Nutri-Score and no saturated fat figure — and saturated fat is one of the UK model's hard requirements. Nyamly cannot score it, so it says No score and Not enough data. NOVA 4 (ultra-processed) still shows, since that classification doesn't need the missing figure.

A blank is not a fifty. It is a blank, and the report keeps it that way.

Nutri-Score: not published. UK score: missing saturated fat, a hard requirement. NOVA 4 stands on its own.

No barcode? Take a photo.

For a plate of food, snap a picture. Nyamly reads back the dishes it sees and a portion guess for each. You confirm the weights before anything is saved.

You get 4 estimates a day, drawn from a weekly pool shared by everyone using Nyamly. The sheet shows your count before you press the button. Every result is marked approximate, because a camera cannot weigh a plate.

How an AI photo calorie estimate works.

Food photo sheet with a pizza, an optional hint field, the note 2 of 3 community scans left, and an Estimate with AI button.

Or just describe it.

Type “grilled chicken plate with rice, and a side salad” and you land on the same review screen: each ingredient, its portion, and how sure the model is. Ingredients matched to a database entry say Matched; the rest say Model estimate only.

A typed estimate always carries a ~ in front of its total. It counts against the same 4 a day.

Review estimate screen: Grilled chicken plate 492 kcal with chicken breast matched to the database and steamed rice as model estimate only; side salad with a low-confidence mixed leaf salad; total 612 kcal; Save to diary.

Two products, side by side.

Scan two. Each row shows both figures per 100 g and a bar for how far apart they are. The colour marks the nutrient, not a verdict. Nyamly leaves the judgment to you.

A row where one product has no published figure says No data and stays blank. It never borrows the other product’s number to fill the gap.

Got a per-serving label? Convert it with the food label calculator.

Compare screen: Fizzy Zero next to Fizzy Original, rows for energy, protein, fat, saturated fat, carbs, fibre and sugar per 100 g, with a note that colour marks the nutrient, not a health judgment.

A diary with a limit.

Log meals against a daily limit you choose. Protein, fat and carbs fill in as you go, and nutrient caps for fibre, sugar, saturated fat and salt sit beneath.

When one meal has no calorie data, the day’s total becomes ~1,200 kcal left with a line saying why. The caps say how much of today’s food actually had fibre data, and that the rest is unknown, not zero.

Today screen showing approximately 1,200 kcal left, limit 2,000 kcal, a note that one meal is missing calorie data so the total is approximate, macro bars, and nutrient caps with the note that 70% of today's calories had fibre data and the rest is unknown, not zero.

Everything you scanned, kept.

The Products tab holds every scan and search, newest first, so you can open a report again without finding the packet.

Scan the same product a few times over a few days and Nyamly suggests a better-scoring swap from things you have already scanned. Scores come from the UK nutrient-profile model, based on your own scan history. Not medical advice.

Products tab: a swap suggestion saying you keep scanning Deluxe Chocolate Yogurt Dessert, scored 13 Bad, try instead Natural Greek Yogurt or Vanilla Skyr Yogurt, scored 82 Excellent; scan history below.