Flour

Best baking substitutes for all-purpose flour

Full ingredient page →

No. 01

bread flour

Sort by protein band: cake ~10%, AP ~11.7%, bread ~12.7%, pastry ~8%, high-gluten ~14%. Within a band: 1:1 by weight. Cross-band: bread <-> AP 1:1; cake/AP 1:1 in pancakes/muffins/quick breads (cornstarch DIY for cakes); pastry/AP 1:1 in crust/biscuits/scones; whole wheat at 50/50 + ~2 tsp water per cup as proportion grows; self-rising/AP needs leavening reconciled (see notes).

Wheat flours swap 1:1 within a protein band; across bands, use the tested cake-flour, bread-flour, whole-wheat, and self-rising conversions instead of a blind 1:1.

No. 02

cake flour

Sort by protein band: cake ~10%, AP ~11.7%, bread ~12.7%, pastry ~8%, high-gluten ~14%. Within a band: 1:1 by weight. Cross-band: bread <-> AP 1:1; cake/AP 1:1 in pancakes/muffins/quick breads (cornstarch DIY for cakes); pastry/AP 1:1 in crust/biscuits/scones; whole wheat at 50/50 + ~2 tsp water per cup as proportion grows; self-rising/AP needs leavening reconciled (see notes).

Wheat flours swap 1:1 within a protein band; across bands, use the tested cake-flour, bread-flour, whole-wheat, and self-rising conversions instead of a blind 1:1.

No. 03

oat flour

No single GF flour swaps 1:1 for wheat. Use: (1) non-yeasted bakes — a tested 1:1 GF blend with xanthan (KAB Measure for Measure, Bob's 1-to-1, Cup4Cup) at 1:1 by weight; (2) yeasted bakes — a tested GF bread blend (KAB GF Bread Flour) at 1:1; (3) DIY: 60-70% structural + 30-40% starch + xanthan by recipe type (see notes). Single GF flours: partial only (~25-30%).

Single gluten-free flours are not 1:1 swaps for wheat; use a tested 1:1 GF blend with built-in xanthan for non-yeasted bakes, a tested GF bread blend for yeasted bakes, or replace at most 25-30% of the wheat with one gluten-free flour plus a binder.

Why these picks

Swaps that preserve structure, moisture, leavening, and browning in baked goods. The ranking favors substitutes for all-purpose flour that preserve structure, absorbency, starch, with verified adjustment notes.

Baking is less forgiving than stovetop cooking; watch hydration, lift, and fat balance after any swap.

Context-ranked swaps

01

bread flour

1 : 1 within a protein band, with cross-band conversions

Wheat flours swap 1:1 within a protein band; across bands, use the tested cake-flour, bread-flour, whole-wheat, and self-rising conversions instead of a blind 1:1.

Read full notes+

Why this works

Within the wheat family, protein level (and for whole-grain or alt-wheat flours, the bran/germ load and gluten quality) drives how the flour behaves. Cake flour at ~10% protein gives the tender, fine crumb cakes are built on; AP at ~11.7% is the workhorse middle ground; bread flour at ~12.7% gives more chew and absorbs more water; pastry flour at ~8% is built for flaky crusts and biscuits; high-gluten flour around ~14% is for bagels, pretzels, and structure-heavy yeast doughs. Same-band swaps are clean. Cross-band swaps are usually workable but not free: bread <-> AP and pastry <-> AP need no other change in most recipes, while cake <-> AP for cakes specifically needs the cornstarch-blend conversion to keep tenderness, and whole-wheat or self-rising swaps need the King-Arthur-tested fixes for water and leavening respectively. Old-world or alt-wheat flours (semolina, durum, spelt, einkorn, kamut, rye, barley) are not interchangeable with AP at full strength because their gluten quality and bran or pentosan loads change the recipe.

Flavor
White wheat-flour swaps taste close to identical. Whole wheat, white whole wheat, and rye taste nuttier and slightly more bitter; spelt, einkorn, and kamut are sweeter and milkier; semolina and durum are pasta-leaning. Pastry, cake, and AP versions of the same recipe all taste neutral.
Texture
Moving up a protein band tightens crumb and adds chew; moving down softens and tenderizes. Cake-flour cakes have a finer, more tender crumb than AP; bread-flour cookies and biscuits are chewier and slightly drier than AP; pastry-flour pie and biscuit crusts are flakier than AP. Whole-grain or alt-wheat swaps reduce rise and produce a denser, coarser crumb in proportion to how much they replace.

Where it fails

High when a recipe is engineered around a protein level: chiffon, sponge, or genoise on cake flour collapses on bread flour; bagels, brioche, or pretzels on bread or high-gluten flour go slack on cake or pastry flour; laminated doughs, croissants, and puff pastry need AP or bread flour. Self-rising into a recipe with baking powder and salt double-leavens unless those are removed. Whole-wheat or alt-wheat above ~30-50% without the King Arthur water bump produces dry, crumbly bakes.

  • Match protein band first: cake flour ~10%, AP ~11.7%, bread flour ~12.7%, pastry flour ~8%, high-gluten flour ~14%. Same-band swaps are 1:1 by weight; cross-band swaps follow the per-direction conversions in this rule rather than a blind 1:1.
  • When swapping bread flour into an AP recipe, expect the dough to absorb slightly more water and add liquid by feel; when swapping AP into a bread recipe, expect the dough to feel slacker. For whole wheat, follow King Arthur's rule of about 2 teaspoons more water per cup of whole wheat as the proportion goes above the 50/50 starting point, and for chilled cookie doughs.

Source: King Arthur Baking: Cake flour vs. all-purpose flour

02

cake flour

1 : 1 within a protein band, with cross-band conversions

Wheat flours swap 1:1 within a protein band; across bands, use the tested cake-flour, bread-flour, whole-wheat, and self-rising conversions instead of a blind 1:1.

Read full notes+

Why this works

Within the wheat family, protein level (and for whole-grain or alt-wheat flours, the bran/germ load and gluten quality) drives how the flour behaves. Cake flour at ~10% protein gives the tender, fine crumb cakes are built on; AP at ~11.7% is the workhorse middle ground; bread flour at ~12.7% gives more chew and absorbs more water; pastry flour at ~8% is built for flaky crusts and biscuits; high-gluten flour around ~14% is for bagels, pretzels, and structure-heavy yeast doughs. Same-band swaps are clean. Cross-band swaps are usually workable but not free: bread <-> AP and pastry <-> AP need no other change in most recipes, while cake <-> AP for cakes specifically needs the cornstarch-blend conversion to keep tenderness, and whole-wheat or self-rising swaps need the King-Arthur-tested fixes for water and leavening respectively. Old-world or alt-wheat flours (semolina, durum, spelt, einkorn, kamut, rye, barley) are not interchangeable with AP at full strength because their gluten quality and bran or pentosan loads change the recipe.

Flavor
White wheat-flour swaps taste close to identical. Whole wheat, white whole wheat, and rye taste nuttier and slightly more bitter; spelt, einkorn, and kamut are sweeter and milkier; semolina and durum are pasta-leaning. Pastry, cake, and AP versions of the same recipe all taste neutral.
Texture
Moving up a protein band tightens crumb and adds chew; moving down softens and tenderizes. Cake-flour cakes have a finer, more tender crumb than AP; bread-flour cookies and biscuits are chewier and slightly drier than AP; pastry-flour pie and biscuit crusts are flakier than AP. Whole-grain or alt-wheat swaps reduce rise and produce a denser, coarser crumb in proportion to how much they replace.

Where it fails

High when a recipe is engineered around a protein level: chiffon, sponge, or genoise on cake flour collapses on bread flour; bagels, brioche, or pretzels on bread or high-gluten flour go slack on cake or pastry flour; laminated doughs, croissants, and puff pastry need AP or bread flour. Self-rising into a recipe with baking powder and salt double-leavens unless those are removed. Whole-wheat or alt-wheat above ~30-50% without the King Arthur water bump produces dry, crumbly bakes.

  • Match protein band first: cake flour ~10%, AP ~11.7%, bread flour ~12.7%, pastry flour ~8%, high-gluten flour ~14%. Same-band swaps are 1:1 by weight; cross-band swaps follow the per-direction conversions in this rule rather than a blind 1:1.
  • When swapping bread flour into an AP recipe, expect the dough to absorb slightly more water and add liquid by feel; when swapping AP into a bread recipe, expect the dough to feel slacker. For whole wheat, follow King Arthur's rule of about 2 teaspoons more water per cup of whole wheat as the proportion goes above the 50/50 starting point, and for chilled cookie doughs.

Source: King Arthur Baking: Cake flour vs. all-purpose flour

03

oat flour

1 : 1 with a tested GF blend, otherwise partial swap + binder

Single gluten-free flours are not 1:1 swaps for wheat; use a tested 1:1 GF blend with built-in xanthan for non-yeasted bakes, a tested GF bread blend for yeasted bakes, or replace at most 25-30% of the wheat with one gluten-free flour plus a binder.

Read full notes+

Why this works

Wheat does two jobs at once: gluten gives structure, stretch, and chew; the wheat endosperm gives a balanced starch/protein bulk that carries the rest of the recipe. Gluten-free flours split those jobs across several ingredients. Rice flour, sorghum, oat, and buckwheat are the structural and bulk flours but cannot stretch on their own; tapioca, potato, cornstarch, and cassava are the bind-and-stretch starches but go gummy or starchy on their own; almond, hazelnut, chestnut, tigernut, and coconut are fat-and-flavor flours with little or no structure; chickpea, teff, quinoa, amaranth, green banana, masa, and corn flour have strong flavors that pull the recipe in a specific direction. The reliably-clean wheat-to-gluten-free path is therefore a tested commercial 1:1 blend that has already been engineered (rice/sorghum/starch/xanthan in tested proportions) for non-yeasted bakes, a tested gluten-free bread flour for yeast bakes, or a homemade 60-70% structural / 30-40% starch blend with xanthan dosed to the recipe type. Single gluten-free flours work cleanly only as partial replacements (typically 25-30% of the total flour) where the remaining wheat or remaining tested blend still carries the structure.

Flavor
Tested 1:1 GF blends taste neutral. Single GF flours: rice and tapioca most neutral; oat/brown rice mildly nutty; sorghum/millet sweet-grainy; buckwheat/teff/amaranth/quinoa earthy and bitter at higher %; corn/masa taste like corn; chickpea is savory-beany (not for sweet bakes); almond/hazelnut/chestnut/tigernut nutty; coconut tastes like coconut; cassava/green banana mild-starchy.
Texture
Tested 1:1 GF blends with xanthan match wheat closely in cookies, brownies, muffins, quick breads, and cakes (slightly tighter/moister crumb). Without a binder, GF batters are runny and doughs sandy. Single starches go gummy alone. Oat/buckwheat absorb more liquid and tighten the crumb. Coconut flour absorbs ~4x its weight — needs extra eggs/liquid. Almond/hazelnut give dense, cookie-like crumb.

Where it fails

High when wheat is replaced 1:1 by a single GF flour or a non-yeasted blend in a yeasted recipe. High in laminated doughs, croissants, and puff pastry (no GF holds layers without specialty fats and binders); popovers and choux (need wheat's gluten film); angel food, sponge, chiffon, and genoise (gluten-stabilized whipped foam). Medium in cookies, brownies, muffins, quick breads, and cakes with a tested 1:1 blend; medium-to-high without xanthan in a homemade blend.

  • For non-yeasted bakes (cookies, brownies, muffins, quick breads, pancakes, pie crust, cakes), use a tested 1:1 gluten-free flour blend that already contains xanthan gum (King Arthur Measure for Measure, Bob's Red Mill 1-to-1, Cup4Cup) at 1:1 by weight or volume. For yeasted bakes (bread, pizza, rolls), use a tested gluten-free bread flour blend (King Arthur Gluten-Free Bread Flour) at 1:1 by weight.
  • When building a blend from scratch or when the chosen 1:1 blend does not already contain a binder, follow King Arthur's xanthan-gum baseline of 1/4 teaspoon per cup of gluten-free flour for cookies, brownies, and quick breads. Scale up to about 1/2 teaspoon per cup for cakes and muffins, and to about 1 teaspoon per cup for pizza dough and yeasted bread. Cap a single recipe around 1 tablespoon of xanthan gum total. Psyllium husk is the more elastic binder option for yeasted GF bread when xanthan alone reads as gummy.

Source: King Arthur Baking: Gluten-free baking guide

04

pastry flour

1 : 1 within a protein band, with cross-band conversions

Wheat flours swap 1:1 within a protein band; across bands, use the tested cake-flour, bread-flour, whole-wheat, and self-rising conversions instead of a blind 1:1.

Read full notes+

Why this works

Within the wheat family, protein level (and for whole-grain or alt-wheat flours, the bran/germ load and gluten quality) drives how the flour behaves. Cake flour at ~10% protein gives the tender, fine crumb cakes are built on; AP at ~11.7% is the workhorse middle ground; bread flour at ~12.7% gives more chew and absorbs more water; pastry flour at ~8% is built for flaky crusts and biscuits; high-gluten flour around ~14% is for bagels, pretzels, and structure-heavy yeast doughs. Same-band swaps are clean. Cross-band swaps are usually workable but not free: bread <-> AP and pastry <-> AP need no other change in most recipes, while cake <-> AP for cakes specifically needs the cornstarch-blend conversion to keep tenderness, and whole-wheat or self-rising swaps need the King-Arthur-tested fixes for water and leavening respectively. Old-world or alt-wheat flours (semolina, durum, spelt, einkorn, kamut, rye, barley) are not interchangeable with AP at full strength because their gluten quality and bran or pentosan loads change the recipe.

Flavor
White wheat-flour swaps taste close to identical. Whole wheat, white whole wheat, and rye taste nuttier and slightly more bitter; spelt, einkorn, and kamut are sweeter and milkier; semolina and durum are pasta-leaning. Pastry, cake, and AP versions of the same recipe all taste neutral.
Texture
Moving up a protein band tightens crumb and adds chew; moving down softens and tenderizes. Cake-flour cakes have a finer, more tender crumb than AP; bread-flour cookies and biscuits are chewier and slightly drier than AP; pastry-flour pie and biscuit crusts are flakier than AP. Whole-grain or alt-wheat swaps reduce rise and produce a denser, coarser crumb in proportion to how much they replace.

Where it fails

High when a recipe is engineered around a protein level: chiffon, sponge, or genoise on cake flour collapses on bread flour; bagels, brioche, or pretzels on bread or high-gluten flour go slack on cake or pastry flour; laminated doughs, croissants, and puff pastry need AP or bread flour. Self-rising into a recipe with baking powder and salt double-leavens unless those are removed. Whole-wheat or alt-wheat above ~30-50% without the King Arthur water bump produces dry, crumbly bakes.

  • Match protein band first: cake flour ~10%, AP ~11.7%, bread flour ~12.7%, pastry flour ~8%, high-gluten flour ~14%. Same-band swaps are 1:1 by weight; cross-band swaps follow the per-direction conversions in this rule rather than a blind 1:1.
  • When swapping bread flour into an AP recipe, expect the dough to absorb slightly more water and add liquid by feel; when swapping AP into a bread recipe, expect the dough to feel slacker. For whole wheat, follow King Arthur's rule of about 2 teaspoons more water per cup of whole wheat as the proportion goes above the 50/50 starting point, and for chilled cookie doughs.

Source: King Arthur Baking: Cake flour vs. all-purpose flour

05

brown rice flour

1 : 1 with a tested GF blend, otherwise partial swap + binder

Single gluten-free flours are not 1:1 swaps for wheat; use a tested 1:1 GF blend with built-in xanthan for non-yeasted bakes, a tested GF bread blend for yeasted bakes, or replace at most 25-30% of the wheat with one gluten-free flour plus a binder.

Read full notes+

Why this works

Wheat does two jobs at once: gluten gives structure, stretch, and chew; the wheat endosperm gives a balanced starch/protein bulk that carries the rest of the recipe. Gluten-free flours split those jobs across several ingredients. Rice flour, sorghum, oat, and buckwheat are the structural and bulk flours but cannot stretch on their own; tapioca, potato, cornstarch, and cassava are the bind-and-stretch starches but go gummy or starchy on their own; almond, hazelnut, chestnut, tigernut, and coconut are fat-and-flavor flours with little or no structure; chickpea, teff, quinoa, amaranth, green banana, masa, and corn flour have strong flavors that pull the recipe in a specific direction. The reliably-clean wheat-to-gluten-free path is therefore a tested commercial 1:1 blend that has already been engineered (rice/sorghum/starch/xanthan in tested proportions) for non-yeasted bakes, a tested gluten-free bread flour for yeast bakes, or a homemade 60-70% structural / 30-40% starch blend with xanthan dosed to the recipe type. Single gluten-free flours work cleanly only as partial replacements (typically 25-30% of the total flour) where the remaining wheat or remaining tested blend still carries the structure.

Flavor
Tested 1:1 GF blends taste neutral. Single GF flours: rice and tapioca most neutral; oat/brown rice mildly nutty; sorghum/millet sweet-grainy; buckwheat/teff/amaranth/quinoa earthy and bitter at higher %; corn/masa taste like corn; chickpea is savory-beany (not for sweet bakes); almond/hazelnut/chestnut/tigernut nutty; coconut tastes like coconut; cassava/green banana mild-starchy.
Texture
Tested 1:1 GF blends with xanthan match wheat closely in cookies, brownies, muffins, quick breads, and cakes (slightly tighter/moister crumb). Without a binder, GF batters are runny and doughs sandy. Single starches go gummy alone. Oat/buckwheat absorb more liquid and tighten the crumb. Coconut flour absorbs ~4x its weight — needs extra eggs/liquid. Almond/hazelnut give dense, cookie-like crumb.

Where it fails

High when wheat is replaced 1:1 by a single GF flour or a non-yeasted blend in a yeasted recipe. High in laminated doughs, croissants, and puff pastry (no GF holds layers without specialty fats and binders); popovers and choux (need wheat's gluten film); angel food, sponge, chiffon, and genoise (gluten-stabilized whipped foam). Medium in cookies, brownies, muffins, quick breads, and cakes with a tested 1:1 blend; medium-to-high without xanthan in a homemade blend.

  • For non-yeasted bakes (cookies, brownies, muffins, quick breads, pancakes, pie crust, cakes), use a tested 1:1 gluten-free flour blend that already contains xanthan gum (King Arthur Measure for Measure, Bob's Red Mill 1-to-1, Cup4Cup) at 1:1 by weight or volume. For yeasted bakes (bread, pizza, rolls), use a tested gluten-free bread flour blend (King Arthur Gluten-Free Bread Flour) at 1:1 by weight.
  • When building a blend from scratch or when the chosen 1:1 blend does not already contain a binder, follow King Arthur's xanthan-gum baseline of 1/4 teaspoon per cup of gluten-free flour for cookies, brownies, and quick breads. Scale up to about 1/2 teaspoon per cup for cakes and muffins, and to about 1 teaspoon per cup for pizza dough and yeasted bread. Cap a single recipe around 1 tablespoon of xanthan gum total. Psyllium husk is the more elastic binder option for yeasted GF bread when xanthan alone reads as gummy.

Source: King Arthur Baking: Gluten-free baking guide

Tools

Use this substitution context in a full recipe or match it against pantry staples.