What records are included?
Every qualifying topic lists all exact FNDDS descriptions in alphabetical order. You can inspect preparation words, additions, and source-serving details before choosing a reference.
Evidence-linked wikiUSDA FNDDS 2021–2023
A food wiki where every answer opens back into the evidence.
Move from a familiar food name to an exact source record, compare genuinely related formulations on one denominator, or trace a listed source ingredient across the meals where USDA uses it. The library preserves the records behind every summary instead of manufacturing one generic answer.
Search the evidence map
Find an exact food, source ingredient, topic, or comparison.
Search stays on this page: it does not create thin, indexable filter URLs. Results link to permanent reference pages with visible source evidence.
How to read this wiki
Specific records first. Interpretation second.
A familiar label such as pizza, soup, yogurt, or fried rice can cover recipes with different ingredients, cooking methods, and portions. The wiki therefore preserves USDA’s exact record names instead of presenting a single generic answer as if every version were identical.
Values are descriptive reference points. They are not measurements of the food in front of you, personalized intake targets, or judgments about whether a meal is “good” or “bad.”
What each topic answers
From a broad food name to the records behind it.
The topic pages are designed for comparison and source discovery, not for manufacturing a precise-looking answer from an imprecise description.
What does the middle look like?
Ten nutrients are summarized with a median and a central 10th-to-90th-percentile interval per 100 grams. The interval shows the middle of the finite record set; smaller topics explicitly warn that a few entries can move the percentile positions.
Why might values differ?
Topic pages count preparation language that actually appears in the source descriptions and list the estimate factors carried by the records. They do not turn those associations into unsupported causal claims.
How do portions change the answer?
Common named household measures and their gram weights remain visible. The wiki separates equal-weight comparisons from source-serving totals because those answer different questions.
Four ways into the evidence
Choose the question you actually have.
The same source corpus can answer different questions only when the denominator and evidence boundary stay visible. These paths are deliberately finite, editorially screened collections—not search-result permutations.
01 · Exact records
3,850 food profiles
Use a precise USDA description, named household measures, Feast’s 21-nutrient panel, the listed source formulation, and its direct FoodData Central citation.
Browse exact profiles → 02 · Collection context107 substantial topics
Study descriptive medians and central ranges without losing the complete catalog—or the small-set caveats—behind them.
Explore source topics → 03 · Formulation evidence623 source ingredients
See where a named USDA formulation input appears, how often it occurs, and what share of listed input weight it represents—without treating batch inputs as a plated recipe.
Trace source ingredients → 04 · Controlled contrasts1,000 record comparisons
Compare two semantically related records per 100 grams, inspect material nutrient deltas, and see the source-description and formulation evidence behind the distinction.
Compare related records →Larger source topics
Start with a substantial collection.
These topics contain the most distinct records in the current corpus. A large record count means more documented variants to inspect; it does not mean the topic is more important.
Mixed dishes
Meat mixed dishes
197 exact records · median 134 kcal per 100 g
Open topic → Pasta, rice & grainsPasta mixed dishes, excludes macaroni and cheese
169 exact records · median 154 kcal per 100 g
Open topic → Meat & poultryPoultry mixed dishes
130 exact records · median 133 kcal per 100 g
Open topic → Meat & poultryChicken, whole pieces
125 exact records · median 206 kcal per 100 g
Open topic → VegetablesOther vegetables and combinations
122 exact records · median 45 kcal per 100 g
Open topic → Breakfast foodsEggs and omelets
116 exact records · median 168 kcal per 100 g
Open topic → BeveragesCoffee
106 exact records · median 32 kcal per 100 g
Open topic → Breads & baked goodsYeast breads
104 exact records · median 277 kcal per 100 g
Open topic → Pasta, rice & grainsRice mixed dishes
99 exact records · median 117 kcal per 100 g
Open topic → Fish & seafoodSeafood mixed dishes
94 exact records · median 134 kcal per 100 g
Open topic → Fish & seafoodFish
89 exact records · median 178 kcal per 100 g
Open topic → Mixed dishesPizza
86 exact records · median 266 kcal per 100 g
Open topic →Browse broadly
Move through familiar food families.
The wiki’s detailed topics use USDA category language. Feast’s broader sections provide a simpler way to begin when you know the kind of food but not the closest database description.
Beans, nuts & protein foodsBrowse profiles →BeveragesBrowse profiles →Breads & baked goodsBrowse profiles →Breakfast foodsBrowse profiles →DairyBrowse profiles →Fish & seafoodBrowse profiles →FruitBrowse profiles →Meat & poultryBrowse profiles →Mixed dishesBrowse profiles →Pasta, rice & grainsBrowse profiles →Sandwiches, burgers & wrapsBrowse profiles →Sauces, fats & condimentsBrowse profiles →Snacks & dessertsBrowse profiles →Soups & stewsBrowse profiles →VegetablesBrowse profiles →
The comparison method
Why every table uses 100 grams first.
Household portions are useful for estimating what someone ate, but two source portions can have different weights. Comparing a cup of one food with a slice, piece, bowl, or unspecified source amount of another can make the food and the denominator change at the same time. A per-100-gram view holds weight constant, so differences between records are easier to inspect.
The topic median is calculated from the sorted reported values and can fall between two observed records. The central interval runs from the 10th to the 90th percentile. It describes the middle 80 percent of the records included on that page; it is not a confidence interval, recommended range, or prediction for a visitor’s meal. The complete observed minimum and maximum remain available in the detail table so unusual records are not hidden.
Source-serving and household-measure values still matter. After finding a close description, open its profile and compare the listed gram weights with the amount on the plate. For a packaged food, a current label may be more specific. For a homemade recipe, ingredient weights and finished yield can be more useful. For an unweighed meal, keep the final number visibly approximate.
Why the wiki avoids a single “healthy” score.
A nutrient table can answer narrow questions about a cited food record. It cannot determine a person’s overall dietary pattern, medical needs, preferences, culture, access, or satisfaction. A record with more of one nutrient may also differ in several others, and equal-weight concentration does not reveal the amount someone will actually eat.
For that reason, Feast reports values, denominators, source descriptions, and uncertainty instead of labeling one record as universally better. If a clinician has given you an individualized nutrition target, use their guidance and the most specific recipe or package information available rather than treating this educational corpus as a treatment tool.