Compare any two foods side by side, per 100 g, using USDA FoodData Central. Check what an ingredient swap does to fat, sodium, sugar or protein before you reformulate.
Values per 100 g as published by USDA FoodData Central.
Each side searches USDA FoodData Central across both generic and branded records and shows the top six matches, with the brand owner where there is one. Once a food is picked on each side, the table lists up to fourteen nutrients - calories, total fat, saturated fat, trans fat, cholesterol, sodium, total carbohydrate, dietary fiber, total sugars, added sugars, protein, calcium, iron and potassium - for both.
All values are per 100 g. There is no portion selector here, deliberately: per 100 g removes serving size from the comparison, so the difference you see belongs to the foods themselves.
The difference is Food 1 minus Food 2. A positive figure means the first food is higher, a negative one that the second is. Where both foods report a nutrient, the higher value is shown in bold.
Because the basis is 100 g of ingredient, the difference is not what your product changes by. To estimate that, scale it by the grams of the ingredient in your recipe: if the two differ by 10 g of fat per 100 g and your batch uses 50 g, the batch changes by 5 g of fat, which is then spread across every serving the batch yields.
A dash means that record does not report the nutrient. A row disappears only when neither food reports it, so a row with one dash is common, and bolding is skipped on those rows because there is nothing to compare against.
The Difference column on such a row treats the missing side as zero, so it simply repeats the other food's value. Read it as "one record has no figure", not as a real difference between the foods.
The most misleading comparisons are between records in different states: raw against cooked, dry against prepared, drained against packed in liquid. Water content alone moves every per-100 g figure, so a swap can look better or worse than it is purely because the two records describe different preparations.
Mixing a branded record with a generic one also mixes kinds of number - a manufacturer's declared, rounded panel against an analysed or calculated value. That is acceptable for a rough look, but for a reformulation decision choose two records of the same kind and the same preparation state.
A comparison tells you the direction and rough size of a change. Whether the change shows on the finished panel depends on the whole recipe, the serving size and rounding, so the reliable test is to rebuild the recipe with the substitute in the label generator and read the result.
If the point of the swap is a claim such as "reduced sodium" or "low fat", check the finished figures against the defined threshold with the claim validator. Claims are assessed on the finished food, not on the ingredient you changed.
Find more foods to compare with Branded Food Search, or list the foods highest in one 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.
Seeing what an ingredient substitution actually costs nutritionally before you reformulate. Swapping one fat or one flour for another moves several nutrients at once, often in opposite directions, and a side-by-side view makes the trade visible rather than leaving it to be discovered on the finished panel.
Per 100 g when comparing the foods themselves, because it removes serving size as a variable. Per serving when you want to know what someone actually gets. Comparisons that quietly switch basis between the two items are the most common way a nutrition comparison misleads.
Not on its own. Nutrition is one axis. Cost, functionality, shelf life, allergen status, supply reliability and how the product tastes all move when an ingredient changes, and a substitution that improves the panel while damaging texture is not usually a win. Use this to inform the decision, not to make it.
Only if it survives rounding. FDA rounding increments are coarse enough that a real difference in a minor nutrient can round to the same declared figure, while a small difference near a boundary can flip a value. Check the change against the rounding rules for the nutrients you care about before assuming the panel will move.
This tool is not tied to one country's rules. What you declare on the label is, and the requirements differ.