Photo scan

Calories from a photo, free: point the camera, get the whole plate.

One photograph is the entire input. Fotocal identifies the dish and the ingredients in it, estimates how much is on the plate, and hands back calories, macros, a health score, any allergens it can see, a short summary from your coach and one realistic swap. Premium opens the panel underneath that: the full micronutrient list, how processed the food is and its glycemic load. Nothing is saved until you say so. Free and no card: 2 AI scans a day; Premium makes them unlimited.

Get it free On Google Play · Android
Real app screen: a scanned chicken and rice bowl, 600 kcal for 450 g, 40 g protein, 70 g carbs and 20 g fat, marked as an 85% AI estimate
How it works

A photo in, a full picture out.

Home-cooked, restaurant, takeaway or last night's leftovers — the input is always the same, and so is the effort.

  1. Photograph the plate

    Whatever is in front of you, however it was made. No good lighting required, no arranging the food, no reference coin next to the bowl.

    Real app screen: the scan camera framing a bowl of chicken, rice, avocado and corn
  2. The AI reads it

    It identifies the dish, separates out the ingredients it can see, and estimates the portion from what is on the plate.

    Real app screen: the detail of a chicken bowl, 'What is in it that you cannot see?': sugar, saturated fat, salt, fibre and cholesterol, then vitamins and minerals as a share of the daily reference
  3. Fix the amounts

    The estimate is a starting point, not a verdict. The portion sheet multiplies the whole plate, ×0.5 for half, ×2 for double, or splits it between two or three people, and the grams, calories and macros follow. Each detected ingredient can also be edited on its own.

    Real app screen: the custom portion sheet over the plate photo, multiplier from ×0.5 to ×3, 450 g, 600 kcal and the macros following it, and the number of people
  4. Correct anything, then save

    Every detected ingredient can be edited, swapped or removed, and you can add what the camera could not see. It is a draft until you save it.

    Real app screen: the home after saving, 682 kcal eaten and 817 left, with breakfast and lunch logged against their targets
What you get

Nine things, from one photograph.

Most trackers stop at a calorie count. The useful part is usually everything after it.

Real app screen, Premium: the full nutrient list of a scanned dish, basics, cholesterol and salt, vitamins and minerals
  • The dish and its ingredients

    Named individually rather than as one blob, so you can see what it thinks it is looking at.

  • Amounts per ingredient

    Read from the plate rather than assumed from an average serving — and each one can be changed, in grams or in units, with minus and plus.

  • Calories and macros

    The headline figure plus the protein, carb and fat split behind it.

  • Two micronutrients free, the rest with Premium

    Every scan reads the full list. The free plan shows the first two, and Premium opens the rest along with fibre, sugar, sodium and saturated fat.

  • A health score

    One number for how the meal stacks up overall, so you can compare two dishes without reading two tables.

  • NOVA processing level

    How processed the food is — a genuinely different question from how many calories it holds, and one that matters over months.

  • Glycemic load

    How much this particular portion is likely to move your blood sugar, based on the amount actually in front of you.

  • Allergens

    Flagged from the ingredients it detects, so a shared or unfamiliar dish is not a guess.

  • A coach summary and one swap

    A plain-language read on the meal, plus a single specific change that would improve it — not a lecture, one suggestion.

The bowl every screen on this page was scanned from: baked chicken on rice with avocado, corn, red onion and yoghurt
In practice

The restaurant plate you cannot look up

No barcode, no label, and no real idea what the kitchen put in it. One photo returns the calories, the score and one specific swap for next time — before the plate has gone cold.

Real app screen: the end of a scan result, how it compares with your week, more protein than your last four scanned meals and 174 kcal over your average, then the exact ingredients with their grams
How it compares

The result knows what else you scanned this week.

Under the nutrients, each scan is set against your own recent meals: more or less protein than your last scanned plates, and how many calories over or under your average scanned meal. Then the exact ingredients it found, each with its grams, ready to edit.

  • Your targets come first: the same meal gets different advice if you are cutting, maintaining or building.
  • It reads the rest of the day. If you are already short on protein at 8pm, the swap will be about protein.
  • It suggests one change, not a rewrite. Something you could realistically do next time, phrased as a choice rather than a rule.
  • Everything you scan feeds the Recommendations tab and your Weekly Report, so the advice gets more specific the longer you log.

Questions about photo scanning

How accurate is a photo estimate?

It is an estimate, and we would rather say so than pretend otherwise. For everyday tracking it is close enough to be useful, and correcting the grams of the main ingredient closes most of the remaining gap. If a meal really matters, weigh the main component and adjust.

What if it gets an ingredient wrong?

Fix it. Every detected item is editable before you save — change it, adjust its grams, replace it, delete it, or add something it missed. Corrections take a couple of seconds and the totals update as you go.

Does it work on restaurant and mixed dishes?

Yes, and that is where it earns its keep. A curry, a stir-fry or a plate of tapas has no barcode and no label, which is exactly the situation database-search apps handle worst.

Is anything logged automatically?

No. A scan produces a draft. Nothing enters your diary, your rings or your totals until you tap save.

Which is the best app for counting calories with photos?

Judge it on four things, not on the demo video: whether it recognises the food you eat, how clearly it says when it is estimating, how fast you can fix grams and ingredients before saving, and what the free plan allows. Fotocal labels every result as an AI estimate with its confidence, lets you change the portion and the ingredients, knows the Spanish dishes, and gives 2 photo scans a day free without a card. Whether that makes it the best for you depends on how often you scan.

Which is the best AI for counting calories?

None of them is right every time, so the best one is the one that shows its work. A photo tells the model what is on the plate but not how much oil went in or how thick the slice is, so every honest app is estimating. What separates them is whether the estimate carries a confidence, whether you can correct it in two taps, and whether it keeps your own history to compare against. Fotocal does the three, and it tells you how sure it is on each result.

Photograph your next meal and see the whole thing.

Fotocal is free to start, with a 5-day free trial on the paid plans.

Get it free On Google Play · Android