Where our data
comes from.
Ingreview's ingredient evaluations combine product-label information, user contributions, open datasets, brand or manufacturer information, third-party product information sources, regulatory references, and our own curated dictionary. This page explains the source categories we use, what each covers, and the attribution obligations we fulfil here.
Reference databases & datasets
Each source category below contributes a distinct layer to Ingreview's ingredient knowledge. No single source is treated as the final word — signals are weighed together and any conflicts are surfaced to the user. For a full explanation of the weighting model, see our Methodology page.
OBF
Open Food Facts & Open Beauty Facts
openfoodfacts.org · openbeautyfacts.orgWhat it is. Open Food Facts (OFF) and Open Beauty Facts (OBF) are collaborative, community-maintained databases of food and cosmetic products from around the world. Products are contributed by volunteers and include ingredient lists, barcode data, packaging photographs, and nutrient information.
How Ingreview uses it. OFF and OBF are the primary source for product lookup by barcode. Ingreview imports a subset of the OFF/OBF data into its own storage and works from that copy, so a scan resolves without waiting on a third-party service. Ingredient lists are parsed there and evaluated against our ingredient knowledge base, and our own corrections and normalisations are applied on top. Ingreview does not publish or hand on the database itself.
License & attribution. The Open Food Facts database is made available under the Open Database License (ODbL) v1.0. Individual contents (product facts) are licensed under the Database Contents License (DbCL) v1.0. Product photographs are not covered by either database licence — they are licensed separately under CC BY-SA 3.0, and wherever Ingreview shows one it carries the credit of the contributor who supplied it. The screens Ingreview shows you are a "Produced Work" under the ODbL. The stored subset we build them from is a Derivative Database, and the ODbL's share-alike terms apply to it — we are building the path to offer our corrections back to the Open Food Facts community, and this page is part of the attribution that goes with it.
ODbL v1.0 Derived subset stored Product catalogProduct labels & product information sources
Labels · user contributions · brand data · third-party product informationWhat it is. Product names, brands, package sizes, barcodes, ingredient lists, nutrition panels, allergen statements, and packaging images may come from product labels, user submissions, brand or manufacturer materials, open datasets, and third-party product information sources.
How Ingreview uses it. Product information helps identify the item a user scanned and, when an ingredient list is available, gives the app text to parse against the ingredient knowledge base. Product data can change when packaging changes, so Ingreview treats it as a product-identification and label-reading layer, not as a permanent guarantee.
Review status. Some product records are complete, some are community-contributed, and some may be incomplete until a user or reviewer adds the missing label details. Ingreview may show lower-confidence or incomplete states rather than pretending every scanned product is fully analysed.
Product labels User contributions Third-party product informationEU CosIng — European Commission Cosmetic Ingredients Database
ec.europa.eu/growth/tools-databases/cosingWhat it is. CosIng is the official cosmetic ingredient database maintained by the European Commission under Regulation (EC) No 1223/2009. It lists INCI names, functions, restrictions, and the regulatory status of cosmetic ingredients in the EU market.
How Ingreview uses it. CosIng is the reference for INCI name normalisation and the primary source for EU regulatory status. When an ingredient is restricted or prohibited under EU cosmetics law (Annexes II–VI), that status is surfaced alongside the ingredient — with the specific annex and concentration limit cited where available.
EU regulation Cosmetics INCI identifiersEFSA — European Food Safety Authority
efsa.europa.euWhat it is. EFSA is the EU agency responsible for scientific assessments of food-related risks. Its published opinions and the EU food additive register (including E-number evaluations) form a comprehensive reference for permitted food additives, their acceptable daily intakes, and any conditions of use.
How Ingreview uses it. EFSA opinions inform the food-additive layer of Ingreview's risk evaluation. Where an E-number has been re-evaluated or an opinion updated, we flag that context so users understand the evidence is current rather than relying solely on historic approval status.
EU regulation Food additives E-numbersSCCS — EU Scientific Committee on Consumer Safety
health.ec.europa.eu/scientific-committees/sccsWhat it is. The SCCS provides independent scientific opinions on the safety of cosmetic ingredients for the European Commission. Its opinions feed directly into CosIng restrictions and inform EU cosmetics regulation amendments.
How Ingreview uses it. Where an ingredient has a specific SCCS opinion — particularly one that led to a regulatory restriction — Ingreview links that context to the ingredient profile. SCCS opinions are referenced as a secondary signal when regulatory status and scientific evidence are in active flux.
EU scientific opinions Cosmetics safetyIngreview's curated ingredient dictionary
Maintained by the Ingreview teamWhat it is. Alongside the public databases above, Ingreview maintains its own ingredient dictionary — a manually reviewed set of entries that adds synonyms, common trade names, parsing rules for complex INCI strings, and allergen vocabulary not yet consistently encoded in the regulatory sources.
How it is maintained. Entries are reviewed against the primary regulatory sources before inclusion. The dictionary is conservative by design: where evidence is limited or contested, entries carry a lower confidence weight rather than a firm positive or negative flag. The dictionary does not contain verdicts — it contains structured evidence summaries, each tied to a cited source.
Wording policy. No entry in this dictionary uses the words "safe", "toxic", or "harmful" as standalone verdicts. Ingreview's wording policy — aligned with our methodology — uses "attention needed", "known allergen", "regulatory concern", and similar qualified language throughout.
Curated Manually reviewed Synonyms & aliasesThis application uses data from Open Food Facts, available at world.openfoodfacts.org and from Open Beauty Facts, available at world.openbeautyfacts.org. The data is made available under the Open Database License (ODbL), its individual contents under the Database Contents License (DbCL), and the product photographs under CC BY-SA 3.0.
Ingreview stores a derived subset of this data and serves results from it. What you see in the app is a "Produced Work" as defined by the ODbL; the stored subset behind it is a Derivative Database, and the ODbL's share-alike terms apply to it. We credit the Open Food Facts community, and we are building the path to offer our corrections back to it. We encourage users to contribute to the project directly.
The alterations we have made. ODbL § 4.6 gives two ways to answer a request from someone who receives a Produced Work: a copy of the whole Derivative Database, or a file of the alterations made to it. We publish the second. It lives at ingreview.app/odbl-alterations.json, it is machine-readable, and it needs no account.
The file is generated from the database rather than written by hand. For every product row we derived from Open Food Facts or Open Beauty Facts it gives the upstream product code and which alterations apply to it, with the rule behind each kind published in full — so anyone holding the upstream data can reproduce the ones that are reproducible. It is not a copy of Open Food Facts and not a copy of our own catalogue. The file also states, inside itself, which kinds of alteration it cannot see; we would rather name the gaps than let the file imply it has none. If something in it looks wrong or missing, tell us.
How these sources combine
Ingreview does not assign a single database the final word on any ingredient. Each source adds a layer of signal, and any disagreements between sources are surfaced explicitly — users see the full picture rather than a single aggregated score.
A full explanation of the evidence-weighting model — including what happens when sources disagree — is on our Methodology page.
Corrections & updates
If you believe a source is missing, incorrectly described, or that our attribution to it is inadequate, please contact us. Accuracy in attribution matters to us both legally and ethically.