Whole-Food Classification Reference

The Diet About Processing, Not Ingredients

Almost every other diet on this site sorts foods by what they are. Vegan excludes animals. Keto excludes carbs. Gluten-free excludes a specific protein. The whole-food diet sorts foods on a different axis entirely: not by what kind of food they are, but by how much something has happened to them between the field and the plate. A potato is a whole food. The same potato in a bag of chips, after being processed, fried, dusted with seasoning, and stabilized with shelf-life ingredients, is not. The category "potato" hasn't changed. The processing has, and the processing is what the diet cares about.

This is a useful reframe because it cuts across most of the usual food debates. Meat is in. Dairy is in. Grains are in. Eggs are in. Legumes are in. Fruit is in. Even sugar and oil are in, as long as they show up as ingredients in home cooking rather than as the structural backbone of an industrial product. The diet doesn't ask anyone to eliminate food groups. It asks them to choose the unprocessed version of whatever they were going to eat anyway, which is a much smaller behavioral change than most diets demand and, somewhat unexpectedly, may produce a larger effect than the size of the change suggests.

The Framework That Actually Makes This Workable

The most useful tool for thinking about whole-food eating is a classification system developed in 2009 by a team led by Carlos Monteiro at the University of São Paulo, now known as NOVA. It sorts every food into one of four groups based on the extent and purpose of the processing it has undergone, and once these four groups are clear, almost every question the diet raises becomes easy to answer.

Group 1 — Unprocessed or minimally processed

Foods in their natural state, or modified only by drying, grinding, freezing, pasteurizing, or simple cooking. Fresh fruit, fresh and frozen vegetables, intact whole grains, dried legumes, raw nuts, eggs, fresh meat and fish, plain milk, plain yogurt, herbs and spices. The processing here removes inedible parts or extends shelf life without adding anything new.

Group 2 — Processed culinary ingredients

Substances extracted from group 1 foods and used as ingredients in home cooking. Oils, butter, lard, salt, sugar, vinegar, honey, flour. These are not foods on their own — they are what you cook group 1 foods with.

Group 3 — Processed foods

Foods made by combining groups 1 and 2, typically with two or three ingredients. Fresh-baked bread with simple ingredients, cheese, canned beans, jarred tomatoes, smoked fish, simple cured meats. These are recognizable as the food they came from, just modified for taste or preservation.

Group 4 — Ultra-processed foods

Industrial formulations made mostly or entirely from substances extracted from foods, plus additives you wouldn't keep in a home kitchen. Soft drinks, packaged snacks, breakfast cereals, instant noodles, frozen meals, sweetened yogurts, "energy" bars, hot dogs and chicken nuggets, most cookies and pastries from the supermarket, "diet" products of every kind. The defining feature is that the food has been disassembled into its component parts and reassembled into something new, usually with stabilizers, emulsifiers, flavorings, and modified ingredients that exist nowhere outside industrial food production.

The whole-food diet, in its honest form, is mostly about moving toward groups 1 and 2 and away from group 4. Group 3 is essentially fine — a piece of cheese, a slice of fresh bread, a can of tomatoes, and a jar of olives are not the problem this diet is solving. The target is the highly engineered industrial food that fills the center aisles of the supermarket and most of the menu at fast-food restaurants. Naming the four groups makes the target precise instead of vague.

Why Ultra-Processed Is the Actual Concern

For most of the past few decades, the nutrition conversation has focused on individual nutrients: calories, fat, sugar, sodium, saturated fat. The whole-food framing argues that the more useful unit of analysis is the food itself, and specifically how much industrial processing has gone into it. The reason this framing has gained traction is a growing body of research suggesting that ultra-processed food affects health outcomes in ways the nutrient label can't fully explain.

The most striking single piece of evidence is a study published in 2019 by Kevin Hall and colleagues at the National Institutes of Health. Hall ran a tightly controlled crossover trial: twenty adults lived in a metabolic ward for four weeks, spending two weeks eating an ultra-processed diet and two weeks eating a minimally processed one. The two diets were carefully matched for calories presented, macronutrients, sugar, salt, fat, and fiber. Participants could eat as much or as little as they wanted at each meal. The result was that people ate roughly 500 calories more per day on the ultra-processed diet and gained weight in the two weeks they were on it. They lost weight and ate less when switched to the unprocessed version, with no instructions to eat differently. The food itself, not the nutrient profile, was driving the difference.

Beyond Hall's study, large observational cohorts have repeatedly linked higher ultra-processed food consumption to higher rates of cardiovascular disease, type 2 diabetes, several cancers, depression, and all-cause mortality, after adjusting for the usual confounders. The French NutriNet-Santé cohort and the Spanish SUN cohort are the most cited. The associations are not enormous in absolute terms but they are consistent across studies and outcomes.

Why this happens is still an open question. The leading hypotheses include hyperpalatability (ultra-processed foods are engineered to override satiety signals), eating speed (these foods require less chewing and reach the stomach faster than satiety hormones can respond), the loss of food structure (the cell walls and physical matrix of whole foods slow absorption in ways processed versions don't), the role of additives on the gut microbiome, and combinations of all of these. The mechanism is contested. The pattern in the data is not.

How to Navigate by the Framework

Once the NOVA framework is in your head, the diet stops being a list of rules and starts being a reflex. Most decisions resolve themselves with a single question: which group is this food in?

Breakfast on a whole-food diet usually shifts away from cereal — even ones marketed as healthy — toward oats cooked from rolled or steel-cut whole grains, eggs in some form, plain yogurt with whole fruit, or simple toast with nut butter and banana. The replacement is less work than it sounds, because most of these are faster to make than they appear. The thing that drops out is the breakfast aisle of the supermarket, which is almost entirely group 4.

Lunch built from group 1 ingredients usually means a salad, a grain bowl, a soup, or leftovers from dinner. The friction is mostly with restaurant lunches and packaged convenience meals, which are dominated by group 4.

Dinner is the easiest meal to keep whole-food because home cooking already tends to live in groups 1 through 3. A piece of meat or fish, vegetables cooked with oil and salt and herbs, a grain like rice or farro, and a simple sauce — none of this requires any change for someone who already cooks.

Snacks are where the diet gets visibly different. The packaged snack aisle is almost entirely group 4, and the replacements are fruit, nuts, hard-boiled eggs, plain yogurt, vegetables with hummus, cheese with bread, or olives. The replacements are not exotic; they are just the foods people stopped reaching for when packaged snacks became cheap and ubiquitous.

The shopping pattern shifts as a result. Whole-food eaters spend most of their time in the perimeter of the supermarket — produce, meat, dairy, eggs, the bakery, the bulk section — and very little in the center aisles, where most of the group 4 products live. The center aisles still have useful items (canned beans, dried pasta, jarred tomatoes, oats, nuts, oils, vinegars) but they are scattered among hundreds of products that aren't whole food, and the cart fills up faster on the perimeter.

Where the Framework Gets Fuzzy

The four-group system is clearer in theory than in the supermarket, and a few categories sit in genuinely ambiguous territory.

Bread. A loaf of fresh-baked sourdough from a bakery, made with flour, water, salt, and starter, is processed (group 3) but very close to whole food in spirit. A typical supermarket sliced sandwich bread, with twenty ingredients including dough conditioners, preservatives, emulsifiers, and added sugar, is group 4. The word "bread" covers both, and the difference matters.

Nut butter. Pure ground peanuts or almonds with maybe a pinch of salt is essentially group 1. The same product with hydrogenated oils, sugar, and stabilizers is group 4. Reading the ingredient list is the difference.

Yogurt. Plain yogurt with two ingredients (milk and cultures) is group 1. Flavored yogurt with sugar, fruit puree concentrate, modified starch, gelatin, and flavorings is group 4. The cup may look identical from the outside.

Pasta. Dried pasta with one or two ingredients (durum wheat, water) is borderline group 3 and is widely accepted on whole-food diets. Filled pasta, "fresh" pasta with stabilizers, or pre-sauced pasta is group 4.

Tofu and plant milks. Plain firm tofu made with soybeans and a coagulant is essentially group 3 and fits the diet. Flavored tofu with sauces and additives is closer to group 4. Plant milks vary enormously — a homemade oat milk is whole-food adjacent; a supermarket oat milk with seventeen ingredients is closer to group 4.

The general rule of thumb is that if a product's ingredient list is short, recognizable, and contains only things you might use in your own kitchen, it is probably acceptable. If the list is long and includes substances you've never heard of, it is probably ultra-processed regardless of marketing claims.

Whole-Food Plant-Based, Briefly

A specific subtype of this diet — promoted by Forks Over Knives, the Esselstyn protocol, and several physician-led programs — combines whole-food eating with veganism. The result is essentially a vegan diet that excludes ultra-processed plant-based foods like vegan ice cream, plant-based meat substitutes, and packaged vegan snacks. It is one of the most studied versions of the diet for cardiovascular disease reversal and has produced impressive results in motivated patients. It is also more demanding than either whole-food eating or vegan eating alone, and most people who try it end up loosening one or the other dimension over time.

If a person is choosing between whole-food and whole-food plant-based, the question is not which is better but which fits the goal. For general improvement and sustainability, the omnivore version of whole-food eating is much easier to maintain and produces most of the benefits. For specific cardiovascular indications under medical guidance, the plant-based version has the strongest evidence.

How This Plays Out Day to Day

The day-to-day shape of whole-food eating is mostly about cooking and shopping habits. The cooking does not have to be elaborate — many whole-food meals are extremely simple — but it has to happen, because the convenience options that fill most modern kitchens are exactly the products the diet asks people to skip. A whole-food eater who tries to live on supermarket convenience meals will fail almost immediately, because those meals are the target of the diet.

The kitchen learns a small set of staples that get used repeatedly. Onions, garlic, olive oil, salt, dried herbs, canned tomatoes, dried beans or canned beans (plain), rice, oats, eggs, plain yogurt, cheese, frozen vegetables, fresh produce, a few proteins. With those, most weeknight meals come together in twenty to forty minutes without recipes. The pattern is closer to ordinary home cooking from a few decades ago than to anything specifically designed.

Restaurants are the dominant friction. Restaurant kitchens depend heavily on ultra-processed components — pre-mixed sauces, dough conditioners, processed oils, packaged stocks, premade marinades, frozen and breaded items — even at restaurants that look high-quality from the outside. The realistic moves are to choose places that cook from scratch (typically smaller, independent, or higher-end restaurants), order simply prepared dishes (grilled fish, roasted chicken, salads, vegetables, simple pastas), and accept that a substantial fraction of restaurant food is closer to group 4 than people realize.

Travel is harder than for most diets. Airport food, road-trip food, and convenience food are essentially defined by ultra-processing. The realistic adjustment is to pack whole-food snacks before leaving — fruit, nuts, hard-boiled eggs, jerky from a brand with a short ingredient list — and to find grocery stores rather than restaurants for at least one meal a day during longer trips.

Cost is genuinely lower than people expect. Whole foods are usually cheaper per calorie and per nutrient than the ultra-processed alternatives, especially when bought in bulk and cooked in batches. The expensive parts of the diet are optional — premium proteins, specialty produce, organic everything — and the foundational version (rice, beans, eggs, frozen vegetables, in-season fruit, basic cuts of meat, plain dairy) is one of the cheapest ways to feed a person well.

Adherence over time is unusually durable for an evidence-based diet, partly because nothing is forbidden in absolute terms and partly because the framing is positive — choose the unprocessed version — rather than restrictive. People who follow the diet for a year or two usually find that ultra-processed foods start to taste off when they encounter them again, the way overly sweet foods taste off after time on a sugar-free diet. The drift, when it happens, is usually toward "mostly whole foods with occasional packaged exceptions," which is itself a meaningfully better pattern than where most people start.

Common Misreadings

"Whole food means raw food." Cooking is fine on this diet. Roasting, steaming, boiling, baking, grilling, and sautéing are all compatible. Whole food is about processing in industrial sense — additives, reformulation, extraction — not about whether heat has been applied.

"All processed food is bad." The framework distinguishes processing levels for a reason. Cheese is processed. Olive oil is processed. Bread is processed. Yogurt is processed. None of these are the target of the diet. The target is the industrial reformulation of group 4, not the modest processing of groups 2 and 3.

"'Natural' on a label means whole food." The word "natural" is essentially unregulated in food marketing and means almost nothing. Many products labeled natural are ultra-processed by any meaningful definition. The ingredient list is the only reliable check.

"Whole-food and Whole Foods Market are the same thing." The store sells plenty of ultra-processed products alongside its produce and bulk bins. The name is a marketing choice, not a guarantee. A whole-food shopper at Whole Foods makes the same decisions a whole-food shopper at any other supermarket makes.

"It's just about avoiding additives." Additives are part of what defines ultra-processing, but the broader concept includes the disassembly and reassembly of food, the loss of food matrix, the engineering for hyperpalatability, and the speed at which processed food is consumed. Removing additives from a group 4 product would not necessarily move it into group 1.

One Habit That Makes the Whole Thing Easier

The single most useful habit on this diet is checking the ingredient list — not the nutrition label, the actual ingredient list — before buying any packaged food for the first time. If the list is short, recognizable, and contains only things you might keep in your own pantry, the food is almost certainly fine. If the list is long, contains chemicals you don't recognize, or is built around extracted components and stabilizers, the food is ultra-processed regardless of how the front of the package is marketed. This single check, applied consistently, accomplishes most of what the diet is asking for, and over a few weeks it becomes automatic. The label is the source of truth. The marketing is not, and the front of the package is the part most carefully designed to mislead.