How it works

What happens after you scan.

Scanning takes a second. This page explains the work behind it: where the product data comes from, how each ingredient is researched and graded, and what a result can and can't tell you.

Read the full methodology

The pipeline

From barcode to result in four steps

It works the same way for food, drinks, cosmetics, personal care and household products. Food-only details, like nutrition, are left out for products you don't eat.

1
Find the product
The barcode number is looked up across Open Food Facts, USDA FoodData Central and a barcode database, and what each one knows is merged into a single set of facts. If none of them has the product, you can add it by photographing the label.
2
Read the ingredient list
The label's ingredient list is split into individual ingredients, so each one can be looked at on its own rather than as one block of text.
3
Research and grade each ingredient
Every ingredient is researched against authoritative sources and graded from beneficial to concern. Each claim carries its citations and the evidence tier behind it. An ingredient that isn't in the research library yet reads Researching while it is researched in the background.
4
Map it to your body
Flagged ingredients are linked to the body systems they touch. Tap an organ on the Body Map to see which ingredients affect it, or tap an ingredient for its full profile.

Product data

Where the product data comes from

Each scan only credits the sources that actually contributed to it.

Open Food Facts

An open, collaborative product database. Used for additive, allergen and processing tags, plus product identity.

USDA FoodData Central

The U.S. government's branded-food dataset. When it has a product, its label ingredients and nutrition numbers take precedence.

Go-UPC

A barcode lookup that fills the gaps with product name, category and image for items the open datasets don't cover yet.

Community submissions

When a product isn't on file, you can add it by photographing its label. It stays private to you until it's reviewed, then joins the catalog marked as user-contributed.

Grading

How each ingredient is graded

Every ingredient gets its own grade and its own evidence. The grade says how it affects you. The tier says how sure the research is.

The grade

From beneficial to concern, scoped to how the ingredient reaches your body and how often you use the product. Whether a product is eaten, applied to skin or used another way is decided once per product. Dose counts too: alcohol is graded by its ABV, and an ingredient at a high dose per serving is graded more cautiously.

Beneficial
Neutral
Mild concern
Concern
Serious concern
Serious harm

The evidence tier

Research is tagged by the kind of study behind it, so you see how strong the evidence is as well as what it says.

Tier 1: Human clinical trialsHighest confidence
Tier 2: Human observational studiesStrong but imperfect
Tier 3: Animal studiesSuggestive
Tier 4: Mechanistic hypothesesExploratory

Honest limits

What you see, and what we don't claim

On every result

A plain-English verdict for the product, with each ingredient graded on its own.
The sources behind every claim, so you can check the research yourself.
A research confidence meter on each graded ingredient: how well the research behind it holds up.
For food and drink, nutrition for the whole container, with a switch to per serving or per 100 g.
The Body Map: which body systems the flagged ingredients touch.
Whether a product holds up at the pace you use it, not just in a single serving.
An unresolved result, and the reason, when the evidence isn't there.

What we don’t claim

A single score or letter grade. It hides the reasoning.
A diagnosis, treatment or medical advice.
Exact dose or how your body will respond. Ingredient order doesn't reveal either.
That a label is complete. Labels can be outdated or differ across markets.
That AI-assisted research is never wrong. That's why the sources stay visible.

Keep reading