understand Blurrix product data sources
Product information has a source and a limit
Blurrix first checks its own product records and can query Open Food Facts, Open Beauty Facts, and Open Products Facts for barcode lookups before creating a scored record.
A barcode starts a lookup, not a guarantee
When you scan a barcode, Blurrix first looks for a matching product in its own product records. A local match can return quickly with the information already associated with that product. When no local record is found, the lookup pipeline can query three external catalogue services in parallel: Open Food Facts, Open Beauty Facts, and Open Products Facts. A returned product may then be classified as food or beauty and passed to the relevant scoring path before it is stored as a product record. If no source returns a product, Blurrix reports that the barcode was not found rather than inventing a label.
These sources are useful because product catalogues cover many countries, brands, and barcode formats, but they are not the package in your hand. Catalogue entries can be incomplete, submitted at different times, translated unevenly, or attached to an older formulation. The product page should therefore be read as a structured record with provenance and limitations, not as proof that every field remains current. A clear nutrition table and ingredient list usually support a more informative food analysis than a record with only a name and image.
Blurrix also keeps product information of its own as products are looked up, enriched, reviewed, or updated. The application does not treat all text as equally reliable. Ingredient handling includes normalization and evidence checks. For example, additive signals used in FoodScore must be traceable to ingredient evidence rather than inferred from a loose substring match. When data is incomplete, the app can use conservative defaults in parts of the scoring model or display an explicit limitation. The point is to avoid making absence look like a positive product feature.
Images and packaging evidence have a separate role. Imported product images can be quality checked and stored when usable, but image retrieval does not block a basic barcode lookup. Packaging analysis may be unavailable when there is no usable front image, the image is too blurry, or the readable evidence is insufficient. In those cases Blurrix records an analysis status instead of displaying a confident packaging interpretation. This distinction is important: a product may have enough label data for one part of its page while lacking evidence for another.
Data also changes over time. The same barcode can later show different packaging, claims, ingredients, nutrition, or score-related evidence. Blurrix records snapshots and can create cautious product-change events when it detects a difference under the same barcode and domain. That is a useful signal to look again, not proof of the date, reason, or commercial significance of a change. Confirm the printed label when the choice matters.
Use the source information as a way to calibrate confidence. If the page matches the package and shows complete ingredients and nutrition, it is a strong starting point for exploration and comparison. If the page is sparse, old-looking, or inconsistent with the item in front of you, do not fill the gaps with assumptions. Scan again, search for the exact product, or rely on the current package. Blurrix organizes available information; it does not replace the manufacturer label or turn a shared catalogue into a permanent record.
Source order matters to the lookup path, not to the truth of a field. A local record may return first because Blurrix already has it. A later external result can carry different or fuller catalogue text. The app stores and processes product information so it can present a usable record, but it cannot certify a manufacturer formulation from a barcode alone. Product names, images, ingredients, and nutrition values should be read as recorded evidence with varying coverage.
If a detail affects the product you are about to buy, compare it with the pack. This is most useful for ingredients, serving information, allergens stated by the manufacturer, and a variant name. A scan that returns no match says only that the configured sources and local records did not return that barcode at that time. It does not establish a score, category, or ingredient list.
Common question
Why can a scanned product differ from its current package?
Catalogue records and formulations can change at different times. Check the package in hand when a detail matters.
Sources
- Open Food Facts Knowledge Base · 2026-07-19
- Open Beauty Facts · 2026-07-19
- Open Products Facts · 2026-07-19
PREVIEW
Full product analysis
Detailed views will be introduced gradually for early-access members.
Get early accessBLURRIX EARLY ACCESS
Less guessing in the aisle. More context when you choose.
A label should answer a question, not create more homework. Blurrix is for the moments when you want to understand what is in your hand without turning a shop into a research project. Keep products, return to them, and compare similar options when you choose again.
Join the early-access list. You will receive an invitation when we open the next places.
Get early access