Gluten-free: the pantry, the rules, and what actually swaps

Gluten is two proteins that, once wet and worked, link into an elastic net. That net is what traps gas in bread, what makes pasta chewy instead of crumbly, and what lets a dough be stretched instead of tearing. Take it out and you have not removed a flavor — you have removed a structure. Everything that follows is about replacing a structure.

This is why swapping flour one-for-one usually fails. It is not that the substitute tastes wrong. It is that nothing is holding the thing together.

What each flour is actually doing

Rice flour is the body: cheap, neutral, and the bulk of most blends. On its own it is gritty and sandy, which is the fault you will recognize first in bad gluten-free baking. Buy it finely milled, and let batters rest so the grains hydrate.

Starches — tapioca, potato, corn are the chew and the stretch. Tapioca in particular gives that slight pull that makes bread feel like bread. Too much and you get gummy; none at all and you get sand held together by hope.

Sorghum, millet, oat, buckwheat are the flavor and the softness. A blend that is only rice and starch tastes of nothing. These are what stop it tasting like a substitute.

Xanthan gum is the net. A quarter teaspoon per cup of flour for cakes, half for bread. It is doing the job gluten did, and it is the ingredient people leave out and then wonder why the loaf collapsed. Too much and it goes slimy and heavy — this is one where more is emphatically not better.

The rules that actually hold

Hydrate more, and rest. Gluten-free flours drink more water and take longer to do it. A batter that looks right immediately is usually too thick once the rice flour has caught up. Twenty minutes of rest fixes more gluten-free baking than any other single change.

Do not be afraid of a wet dough. Gluten-free bread dough should look like thick batter, not like a kneadable ball. There is no gluten to develop, so there is nothing to gain from a stiff dough and everything to lose.

Weigh things. Flour blends vary enormously in density, and cup measures of rice flour and tapioca starch are not interchangeable by volume even when the recipe pretends they are.

Do not knead. There is nothing to develop. Mix until combined and stop.

The four that hide

These are the ones that catch people who have been doing this for years.

- Soy sauce is made with wheat. Tamari usually is not — check the bottle, because "usually" is doing real work in that sentence. - Oats are naturally gluten-free and are commonly cross-contaminated in milling, which is why certified oats exist and cost more. - Stock, bouillon and gravy granules are often thickened with wheat flour. Check the label rather than assume. - Beer, and anything deglazed with it. Also malt vinegar, and anything called "malt".

And one more that is not an ingredient: shared equipment. A separate set of the cheap things — boards, a toaster, deli containers — is easier than policing one set, and most of it is on the restaurant supply list. A toaster, a fryer, a flour-dusted board. If it matters medically, it matters here more than any label does.

Bread, and why it is its own problem

Everything above applies to cakes, batters and pastry. Bread is harder, because bread is the one thing where the structure has to hold gas for an hour and then survive losing it.

Psyllium is not a powder you stir in and use. It has to hydrate and gel before the dough is workable, and that takes time sitting still — the dough goes from a slack, tacky mess to something that holds together and can be shaped. Skipping that wait is the most common way mine have failed, and it is invisible: the dough looks wrong, you add flour to fix it, and now the loaf is dense for a reason that had nothing to do with the flour.

The vessel you proof in decides the shape. The one you bake in does not. Gluten-free dough has no structure of its own to hold a form, so it takes the shape of whatever it rises in and then keeps it. A round loaf comes from proofing in something round and straight-sided, not from baking in something round. This is the opposite of wheat bread, where a banneton is a convenience and the oven spring does the rest.

It tells you it is ready by holding its shape, not by feel and not by the clock. The poke test is a gluten test — it reads the elastic net, and there is no net here. What you are looking for instead is a dough that has risen and stays domed rather than slumping when it is disturbed. A clock cannot tell you that, because how fast it gets there depends on the room.

Gluten-free flours hydrate slower and stale faster. Slower going in, which is why resting matters at every stage; faster coming out, which is why a gluten-free loaf is noticeably worse the next day when the same wheat loaf would be fine. Bake it the day you want it, or slice and freeze it the day it is baked rather than the day it goes stale.

Scale, tin size and timings belong to a recipe rather than to this page, because they change with the pot. They are in the bread recipe.

Choosing a blend

A bagged 1-to-1 blend is the easier route and gives you less control. It already contains xanthan and a fixed balance of starch to flour, which is what lets it stand in for wheat flour by volume. That is the whole appeal, and it is also the limit: the structure is decided for you.

A scratch blend lets you tune structure and flavor separately. A flavor flour such as sorghum or millet, a starch such as tapioca, then psyllium and xanthan for structure. Four decisions instead of one, and each of them moves something different.

They are not interchangeable. A recipe written for a 1-to-1 will not behave in a scratch blend and the reverse is worse, because a scratch recipe already contains its own gum and adding a blend that contains more gives you something closer to rubber than bread.

Bread, in practice

Grind psyllium fresh from whole husk. Pre-ground powder varies enormously in density, so the same weight is a different amount of husk from one bag to the next, and the structural ingredient is the last one you want to be guessing at.

Bake to internal temperature, not to a time. 208 to 210°F in the center. A gluten-free loaf can be deep brown outside and still wet through the middle, and color is not evidence.

A paddle, not a dough hook. A hook works by wrapping gluten around itself. There is none here, so a hook pushes the dough around the bowl while a paddle forces water into the starches, which is the actual job.

Scaling a loaf up is not linear. More dough has to drive more moisture out through the same amount of crust in the same time, so a recipe that works at one size gives a damp center at the next. Bigger loaves need their own weights and their own timings, not a multiplier.

Worked through end to end in the gluten-free Dutch oven bread, which carries two tested sizes rather than one scaled one, for exactly that reason.