Calorie counter with a barcode scanner built in

Why scanning changes whether counting works

Manual entry means searching a name, choosing between a dozen near-identical results, and converting a portion. Perhaps forty seconds per item. Five items a day is over three minutes of fiddly work, every day, forever.

Scanning is two seconds and involves no decisions. The difference is not convenience, it is whether you are still doing it in a fortnight.

It also removes the largest source of error, which is not arithmetic but selection. When you search "cheddar" and pick the fourth result, you have guessed. A barcode identifies one specific product with no ambiguity at all.

What happens after the scan

The scan is the easy part. What a counter does next is what separates a good one:

It defaults to a real portion. Databases store per 100 g. A counter that logs 100 g of everything is wrong on nearly every entry. Ours defaults to the manufacturer's serving, shows what it weighs, and recalculates as you adjust.

It puts the item in a meal. Breakfast, lunch, dinner or snack. This seems cosmetic and is not: meal-level patterns are where the useful insight lives. Most people discover their problem is a specific slot, not their overall discipline.

It shows the running total against a target. A number with no context is just a number. What matters is how much is left.

It remembers. Most people eat the same forty foods. A counter that surfaces your recent items turns a repeat breakfast into one tap.

The logging mistakes that make the total wrong

Logging the serving instead of the amount

Accepting "1 serving" without checking what a serving weighs. Cereal boxes commonly declare 30 g; an actual bowl is often 60–80 g. That single habit can put a breakfast out by 150 calories.

Forgetting the cooking fat

The most under-recorded item in home cooking. A tablespoon of oil is about 120 calories and it is in the food, not the pan. If you cook, log it.

Skipping drinks

Juice, a latte, beer, a mixer. Easy to consume, easy to forget, and often several hundred calories. Alcohol carries about 7 calories per gram, second only to fat.

Not logging the bad days

The most damaging one. Skipping a day you would rather not record turns the diary from a record into a highlight reel — and because the skipped days are systematically the heavy ones, the average it reports is not merely imprecise, it is biased downward. A logged bad day is information. An unlogged one is self-deception with a chart attached.

Counting a weekly average as a daily one

Intake swings for entirely ordinary reasons. Judge the seven-day average; the daily figure is too noisy to act on.

How much precision is worth it

Calorie counting is a steering tool, not a measurement. Labels are permitted about 20% deviation, absorption varies between people, and the energy cost of digestion differs by food. Anyone reporting their intake to the nearest calorie is reporting false precision.

What counting is genuinely good at is showing you patterns you could not otherwise see: that lunch is consistently larger than you thought, that weekends undo the week, that a particular snack accounts for more than the meal it follows. Those findings survive a 20% error easily.

So scan what scans, estimate the rest, log everything, and look at the week.

Setting a target that is not nonsense

A target should come from your own maintenance level, not from a round number. Estimate maintenance with the Mifflin-St Jeor equation and an activity factor, then subtract 300–500 calories for a loss of roughly 0.25–0.5 kg a week.

Larger deficits are not better. Below about 1,500 calories for men or 1,200 for women, hitting protein and micronutrient needs becomes difficult, adherence collapses, and more of the weight lost is muscle. A deficit you can hold for twelve weeks beats an aggressive one you abandon in two.

Protein is worth setting explicitly — roughly 1.6 g per kg of body weight if you are losing weight, which protects muscle and is the most satiating macronutrient per calorie. The macro tracker page goes into the rest.

Start with one scan — that one is free.