Nutrition Facts Search
Search any food in USDA FoodData Central and see its full nutrition facts per 100 g and per common portion, with %DV against the 2016 FDA reference values.
Data: USDA FoodData Central. Results are cached in your browser.
Search any food in USDA FoodData Central and see its full nutrition facts per 100 g and per common portion, with %DV against the 2016 FDA reference values.
Data: USDA FoodData Central. Results are cached in your browser.
This search queries USDA FoodData Central and is restricted to its generic food datasets - Foundation Foods, SR Legacy and FNDDS - rather than packaged retail products. Each result shows its data type and food category, so you can see which dataset a record belongs to before you open it.
An open record also shows its FDC ID, the identifier for that exact entry. It is worth noting alongside any figure you carry into a formulation, because "chicken breast" matches many records and the ID is the only thing that says which one you used.
Every value starts per 100 g. Choosing a portion from the Amount menu - "1 cup, chopped" or "1 medium", with its gram weight - multiplies each nutrient by that weight divided by 100. The portions are the household measures USDA attached to that record, and only the ones with a gram weight are offered.
Those portions describe the food, not your product. The serving size on a Nutrition Facts panel is derived from the reference amount for the product category, not from whichever measure a database happens to list - see the RACC lookup for that figure.
Values are displayed to the nearest calorie and to one decimal place for everything else. That is display rounding, not the FDA rounding rules a label applies.
One search term usually returns several records that look alike: the same food in different datasets, raw and cooked, drained and undrained, with and without salt, or as different cuts and varieties. They are not duplicates. They describe different things, and their numbers differ accordingly.
Pick the record that matches your ingredient as it goes into the recipe. If your recipe weighs dry pasta, a cooked-pasta record will understate almost everything per 100 g, because cooking adds water. The reverse error - a raw record for an ingredient you weigh after cooking - overstates it.
The percentage is calculated against the 2016 FDA reference values for adults and children 4 and over, for whatever amount is selected, and rounded to a whole percent. It matches your panel only if the selected amount is your serving and the record matches your ingredient.
Rows appear only when the record reports that nutrient. A missing added sugars or vitamin D row means the record holds no value for it, not that the food contains none - and that distinction matters if you are about to sum ingredients into a panel.
A database value describes the samples USDA measured or calculated, not your batch. For a single-ingredient product with a well-matched record it may be most of the work; for anything made from a recipe it is one input among several, and the panel has to be built from the whole formulation by weight.
That is what the label generator does: it sums ingredient values by the weights you use, then applies serving size, rounding and %DV for the finished product. Copying one ingredient's record onto a multi-ingredient label skips every one of those steps.
To compare foods side by side, use Food Comparison, or rank them by a single nutrient with Nutrient Ranking.
Reference information, not legal advice. Requirements change and turn on facts specific to your product - confirm against the current regulation before relying on it.
From published food composition data, most of it traceable to USDA FoodData Central. That is a database of analysed and calculated values for generic foods, which makes it the right starting point for formulation and for a first estimate of a recipe.
Yes, with care. FDA accepts values computed from a recognised nutrient database rather than requiring laboratory analysis. The responsibility for accuracy stays with you, so the database entry has to genuinely match your ingredient: the right cut, the right form, the right moisture state, and the right supplier where it matters.
When raw material varies significantly batch to batch, when your process changes composition in ways a database cannot model, when a nutrient sits close to a claim threshold or a rounding boundary, or when a customer or regulator asks for analytical evidence. Anywhere a small error would change what the label says, testing is the cheaper option.
Because they are describing different samples. Variety, season, soil, feed, ripeness, cut and preparation all move nutrient values, and databases report a mean from a finite sample. A spread of ten to twenty percent between sources for a fresh product is normal rather than an error in either.
This tool is not tied to one country's rules. What you declare on the label is, and the requirements differ.