From Barcode to Breakfast: Where Moow’s Food Data Comes From
A barcode identifies a product record; it does not guarantee the record is current or the serving matches your bowl. Moow keeps source, serving, and user corrections separate on purpose.
By Moow Product ·
- Moow uses USDA FoodData Central as the initial canonical source.
- Nutrients must be scaled from the source basis to the serving the person actually chose.
- Custom foods and proposed corrections remain separate from verified catalog records.
The barcode is a lookup key, not the nutrition
When Moow reads a package barcode, the digits are sent to the food lookup endpoint. The camera is detecting the code; it is not photographing the package for nutrition analysis, and the barcode flow does not save raw photo or video. A match returns a catalog identity with brand, market, serving, source, and quality information. A miss sends the person to search or manual entry.
A successful match still needs review. Manufacturers reformulate products, package sizes change, regional versions differ, and multiple serving descriptions can exist. The code on the box identifies what record to inspect; the current packaging remains the practical source for allergens and the most recent manufacturer-specific instructions. Scanning should shorten navigation, not suppress verification.
USDA data enters through a controlled pipeline
Moow uses USDA FoodData Central as its initial source of truth. Source releases are imported into versioned raw tables, normalized, then published into a separate application catalog. The pipeline records release information, row counts, checksums, and validation results so a failed or incomplete batch is not silently presented as current food data.
The development environment currently has a published Foundation Foods subset. The broader filtered artifact includes Foundation foods, FNDDS foods, and selected active US branded records with valid barcodes, but live branded publication depends on the validated database import. Barcode success cases can use fixtures during development. That boundary should remain visible until the production catalog is actually loaded.
- Foundation foods: core ingredients with detailed source data.
- FNDDS foods: foods commonly reported in dietary intake surveys.
- Branded foods: product-specific records that require barcode and recency checks.
- User-created foods: private records that are never silently promoted to verified catalog data.
The serving choice is where the math happens
Food nutrients may be expressed per 100 grams, per 100 milliliters, or per serving. After a person selects a food, Moow loads its available servings and primary nutrients, then scales calories, protein, carbohydrate, and fat to the chosen quantity. If no source serving exists, the interface can offer a calculated 100-gram fallback rather than invent a household measure.
That calculation is only as good as the chosen serving. “One bowl” is not automatically the same as the source serving, and draining, cooking, or dilution can change the relationship between weight and nutrients. The normal meal form opens before saving so the user can review meal type, quantity, and nutrition. Historical meals keep their food and serving links for future traceability.
Verified source data and a wrong serving still produce a wrong meal total. Review both.
Corrections are proposals, not instant edits
People can save a manual meal as a reusable private custom food. That record carries a user-created quality label and remains owned by the account. For a canonical food, a user can instead propose a correction with a field, value, and optional evidence. The proposal does not mutate the live catalog; an authorized reviewer can accept or reject it, with an audit record of the decision.
Even an accepted proposal waits for a separately controlled catalog publication step. This separation is deliberate: community input can improve discovery without letting an unreviewed edit rewrite nutrition for everyone. The product should show source dates and quality states in plain language so users understand whether they are looking at USDA data, a private custom item, or a pending suggestion.
Sources
For allergies, intolerances, medical diets, or recalled products, verify the physical package and manufacturer information rather than relying only on a database match. Do not consume a product when the identity, ingredients, or label are uncertain.