Do You Need a Barcode Scanner to Count Calories?
No — unless most of what you eat is still wearing its packet when you eat it. A barcode scanner is the fastest and most exact way to log packaged food, and it is no help at all with anything cooked, plated or served to you. The question is how much of a normal week of your food still has a barcode on it when it reaches your mouth. For some people that is most of it. For most people who cook, it is the smaller half.
What a scanner is actually doing
A barcode carries no nutrition information. It is a number — the same EAN or UPC the supermarket till reads — and the app uses it as a lookup key. When a scan works, what comes back is an entry a person typed in at some point: a manufacturer, a volunteer, or another user of the same app. The camera is reading a number, not a label.
Many apps look that number up in Open Food Facts, which describes itself as a collaborative, free and open database of food products "made by everyone, for everyone", run as a non-profit by volunteers; it listed about 4.7 million products when we checked on 2026-09-10. Others keep their own crowd-sourced catalogues, where the entry behind a barcode came from a previous user and may never have been checked.
Three things make scanning excellent for the food it covers. It is fast — two seconds against a minute of typing and choosing between near-identical results. It involves no estimation, because the pack's own declared values come back rather than a guess at a similar product. And it is repeatable: the bar you scan on Tuesday returns the same entry on Thursday, and day-to-day consistency is what makes a diary readable a month later.
Disclosure: Olivka is our product and it appears further down. It has no barcode scanner. It reads a nutrition label from a photo instead, which is a different tool with different limits, and this post says where each one wins rather than pretending they are the same thing.
Where the number behind the barcode goes wrong
The lookup is exact. The entry it finds is not guaranteed to be. A crowd-sourced entry can be typed in wrong, be years out of date after a reformulation, or be one of several duplicates with different values for the same pack. A barcode identifies a product line, not a batch and not a country: the same code can be sold in several markets with a locally adjusted recipe, and the database holds whichever version somebody entered first. If a scan returns something that looks off, check the entry against the pack once before trusting it for a month.
Then there is the label itself. In the United States, 21 CFR 101.9(g) deems a food misbranded when its calories, sugars, total fat, saturated fat, cholesterol or sodium exceed the declared value by more than 20 percent — while "reasonable deficiencies" of those same nutrients under the labelled amount are acceptable. So a US label may understate calories by up to a fifth and still comply, and the rule is written so that the tolerated error runs in the direction that flatters the product. The European Commission's guidance on Regulation 1169/2011, dated December 2012, is symmetrical but not small: ±20 percent for carbohydrate, sugars, protein, fibre and fat declared between 10 and 40 g per 100 g, and fixed gram tolerances outside that band — ±2 g below 10 g and ±8 g above 40 g for carbohydrate, sugars, protein and fibre; ±1.5 g and ±8 g for fat — with measurement uncertainty already included.
None of that argues against scanning. It argues against the idea that a scanned number is exact: the scanner fetches the label's figure perfectly, and the label's figure was allowed to be off before you started. Our accuracy page goes through why that floor applies to every logging method, not only this one.
What the scanner cannot see
A scanner answers one question — what is in this pack, per 100 g or per serving — and nothing else. Everything a barcode does not cover falls back on you.
| What you ate | Scannable? | What you still have to do |
|---|---|---|
| A protein bar, the whole thing | Yes | Nothing — this is the ideal case |
| Half a 150 g bag of crisps | Yes, for the bag | Weigh or estimate the half you ate |
| Oats with milk and a banana | Oats and milk, yes | Weigh both; the banana has no barcode |
| A curry you cooked from scratch | The rice and the tin of tomatoes | Everything else — the dish has no entry |
| Lunch in a restaurant | No | Estimate the whole plate |
| Loose fruit, vegetables, bakery bread | No | Search or estimate |
| Sunday meal prep, five boxes | The ingredients, once | Divide the batch; the box has no code |
Two of those rows catch people out more than the rest. The partial pack: the app logs the entry and then asks how much, and the honest answer to how much of a bag of crisps you ate is usually a guess — so the estimation the scanner promised to remove walks back in through the portion field. And home cooking: the components scan, the meal does not, so a scanner-first app pushes you toward recipe builders and weighing every ingredient, which is accurate and which almost nobody sustains past the second week.
How much of your food actually has a barcode?
There is no honest general figure — it depends on how you eat — but you can measure your own in a minute. Write down yesterday, every item, and mark each one that reached your mouth still in its packet, in roughly the printed amount. That is your scannable share.
- If you buy lunch as a meal deal, snack from wrappers and eat a ready meal or two a week, most lines get a mark, a scanner covers the bulk of your intake, and the rest is a few estimates.
- If you cook most evenings, the marks land on cereal, milk, yoghurt and the occasional bar — and every dinner, the biggest meal of the day, is unmarked. The scanner handles the small half by calories.
- If you eat out or order in several times a week, almost nothing gets a mark. The scanner is close to irrelevant, and whatever estimates a plate is the tool doing the work.
The share by count and the share by calories are different things. Snacks scan; dinners do not. A scanner that covers most of your entries can still cover well under half of your energy, so count the calories in the marked lines, not the lines.
Which apps scan for free (checked 2026-09-10)
Everything below was read from the vendor's own site on 2026-09-10. Which tier the scanner sits in is the fact that moves most often in this category, so if you are reading this months later, look again.
MyFitnessPal's Premium page lists "barcode scan and meal scan" as included with Premium, shown at $79.99 a year on the page we saw. It was not always so: free accounts had the scanner for years until MyFitnessPal moved it behind Premium on 1 October 2022, the single change most responsible for people searching for a free one.
Cronometer's homepage says "scan the barcode of any product — for free", and describes the free version as feature-rich alongside a paid Gold tier.
YAZIO's site describes tracking "manually or by barcode" and the app as freemium, but the public pages we could open do not say which tier the scanner sits in. Assume nothing; check inside the app before you enter a week of meals.
Fitatu, the Polish app, lists a barcode scanner among its features and a database it says covers 91 percent of products on the Polish market; its Premium page names ads, recipes, web access, fasting schedules and fitness-app syncing as the paid extras, not the scanner. If paying is the objection rather than the scanner, we have a separate piece on what "free" means in each of these apps.
What each free tier includes beyond the scanner is compared in the best free calorie counting apps.
The alternative: photograph the label
There is a second way to get a pack's own numbers into a diary, and it needs no database at all. Photograph the nutrition panel. A vision model reads the table — energy, fat, carbohydrate, protein, per 100 g or per serving — you say how much you ate, and the figures logged are the label's own, not a crowd entry for a possibly different recipe.
The arithmetic that follows is the whole job, and it is small. The label says 380 kcal per 100 g; you ate 40 g; that is 152 kcal. It is what a scanner does after a successful lookup, minus the lookup — so it works on packs no database has seen: a bakery label, an import, an own-brand launched last week.
It loses on speed for the item you eat every day. A scan needs no framing; a photo of a panel needs light and a steady hand. If the same bar is in your bag every afternoon, scanning it is faster than photographing it — though logging it by name from your own history is faster than either.
Olivka works the second way. Send a photo of the label in WhatsApp, Messenger, Telegram or Instagram, tell it the portion, and the meal card comes back with the label's values and a Confirm or Edit button. For the plate that has no label — the curry, the restaurant lunch — you send a photo of the food instead and it estimates — a different and less exact process, described in our piece on tracking from a photo. Corrections replace the earlier entry rather than adding to it. There is no scanner. The first meals are free and after that it is a paid subscription, with the plan shown at the end of onboarding; we do not quote a price the site itself does not publish. How the estimating half compares with other tools is in our overview of AI calorie trackers.
How to decide
- Most of your food comes in packets and you eat the printed serving: pick an app with a scanner, and check which tier it sits in before you start. Speed and repeatability will beat everything else on this page.
- Most of your food is cooked at home or eaten out: a scanner solves the smaller half. Choose by how well the tool handles a plate, and treat label reading — by scan or by photo — as the secondary feature.
- Both, honestly: many people want a scanner for the cupboard and something that can look at a plate for everything else. Two tools is fine, but two diaries do not add up — decide which one holds the day's total and type the other's numbers into it.
One habit makes either choice work better: log what you ate, not what the pack contained. The scan is exact about the bag; only you know about the half still in the cupboard, and that gap is where most "the app says I'm in deficit but nothing is changing" stories live.
Frequently asked questions
Do I need a barcode scanner to count calories?
Only if most of what you eat is still in its packet when you eat it. A scanner is the fastest and most exact way to log packaged food, because it fetches the pack's own declared values with no estimating. It does nothing for cooked meals, restaurant food, loose produce or the portion of a pack you actually ate, and for most people who cook those are the larger half by calories. Write down yesterday, mark what was still in its wrapper, and you have your answer in a minute.
How accurate is a barcode scanner for calories?
The lookup is exact; the numbers it finds are not. The entry may be crowd-sourced, out of date or for another market's recipe, so check it against the pack once. And the label itself is allowed to be off: in the US, 21 CFR 101.9(g) only misbrands a food when calories, sugars, fat or sodium exceed the declared value by more than 20 percent, while reasonable shortfalls are acceptable; the EU's guidance allows ±20 percent in the 10–40 g per 100 g band, with fixed gram tolerances outside it. A scan gives you the label's figure perfectly, and the label's figure was never exact.
Which calorie apps have a free barcode scanner?
As of 2026-09-10, reading the vendors' own sites: Cronometer says its barcode scanner is free. MyFitnessPal lists barcode scan under Premium, shown at $79.99 a year — it was free until 1 October 2022. Fitatu lists a scanner among its features and its Premium page does not name it among the paid extras. YAZIO's public pages describe barcode logging but do not say which tier it is in. This is the fact that changes most often in this category, so check the current pricing page before you commit a month of meals.
Can I log packaged food without a barcode scanner?
Yes. Photograph the nutrition panel instead. A vision model reads the table, you state how much you ate, and the numbers logged are the label's own — 380 kcal per 100 g times 40 g is 152 kcal, which is all a scanner does after its lookup anyway. It works on packs no database has seen, such as bakery labels or imports, and it is slower than a scan for the bar you eat every day. Olivka does this in WhatsApp, Messenger, Telegram or Instagram; it has no scanner, the first meals are free, and after that it is a paid subscription.