Scan a label right here

Upload a photo of any ingredient list or type it in. You will get a health score, the additives worth knowing about, what the sugar actually is, and a less processed alternative.

Or type it in

Food label scanner apps have become genuinely popular, and for good reason. Ingredient lists are written for regulators, not shoppers. A list of thirty ingredients including four different names for sugar and six additive numbers is technically informative and practically useless.

A scanner does the translation. But like any tool, it is worth understanding what it is actually measuring, because a score of 62 tells you nothing on its own.

What a Scanner Actually Does

There are two ways these apps work, and they produce different results.

Barcode scanning looks up the product in a database — often a crowd-sourced one like Open Food Facts — and returns whatever data has been entered. It is fast, but it depends entirely on the database having accurate information for that specific product, including its country variant.

Label reading uses optical character recognition and language models to read the actual ingredients and nutrition panel from a photo. It works on any product, including ones with no barcode and ones the database has never seen, but it depends on the photo being legible.

The second approach is more robust for anything unusual — a farmers market product, a foreign import, or a shop's own brand that has been reformulated. Our guide to how ingredient scoring works covers the database approach in more detail.

What the Score Is Actually Measuring

Most health scores are a weighted combination of a few factors. Understanding them lets you read a score sensibly.

  • Sugar content. Usually weighted heavily, and often counting all sugars including those naturally present in fruit or dairy.
  • Salt. Sodium is weighted heavily in most scoring systems because of the strong evidence linking it to blood pressure.
  • Saturated fat. Typically penalised, though the evidence is more nuanced than most scores suggest.
  • Fibre and protein. Usually rewarded, because both improve satiety and metabolic response.
  • Additives. Some systems penalise the number of additives; others weight specific ones. The count-based approach is crude.
  • Processing level. Some systems use NOVA classification, which groups foods by how processed they are rather than by nutrient content.
Discover more
Get Produce Boxes
fresh produce
Produce

Where Scanners Get It Wrong

Natural sugar counts the same as added sugar. A fruit yoghurt and a confectionery bar may both score poorly on sugar, despite being nutritionally very different foods. Some systems distinguish added sugars; many do not.

Additive count is a blunt instrument. A product with six harmless additives scores worse than one with three, even if the three are more concerning. Counting is not the same as assessing.

Processing level and nutrition are different questions. A fortified breakfast cereal might be ultra-processed by NOVA classification but nutritionally reasonable. Conversely, some minimally processed foods are nutritionally poor.

Portion size assumptions. Nutrition panels are per 100 grams, but nobody eats 100 grams of most things. A score that ignores realistic serving size can mislead.

What They Are Genuinely Good At

Despite those limits, a scanner does several things really well.

  • Spotting hidden sugar. Manufacturers split sugar across many names so that no single one appears first on the list. A scanner catches all of them. Our guide to hidden sugar names lists over sixty.
  • Decoding additive numbers. Turning "160b, 330, 322" into plain English is genuinely useful, particularly for anyone avoiding specific additives.
  • Flagging allergens. Cross-checking an ingredient list against your own allergen list is exactly the sort of tedious task a machine should do.
  • Comparing similar products. Two pasta sauces side by side, scored the same way, is a much fairer comparison than reading two panels.

Reading a Score Sensibly

The most useful way to use a scanner is comparatively rather than absolutely. Do not ask "is 62 good?". Ask "is 62 better or worse than the alternative on the shelf?" That question the tool can answer well.

It is also worth reading the breakdown rather than just the number. If a product scores badly because of sugar, that is different from scoring badly because of additives, and the right response differs.

And remember the whole-food rule. A bag of carrots is not going to have a barcode or an ingredient list, and it does not need one. The most reliably healthy things in a supermarket are the ones no scanner can score.

The Short Version

A label scanner translates an ingredient list into something readable and scores the product on sugar, salt, fat, fibre and additives. It is excellent at catching hidden sugars and decoding additive numbers, and poor at distinguishing natural from added sugar or assessing additives by anything other than count. Use it to compare products rather than to judge them in isolation.

Frequently Asked Questions

Either by looking up a barcode in a product database, or by reading the label directly from a photo using optical character recognition and language models. The photo approach works on any product, including ones with no barcode.