Allulose (D-allulose, also called D-psicose) is a "rare sugar" — a true sugar molecule that occurs naturally in small amounts in foods like figs, raisins, and maple syrup. It tastes about 70% as sweet as table sugar and behaves like sugar in cooking — browning, dissolving, depressing freezing point — but the human body absorbs it and then excretes most of it unchanged, so it contributes roughly 0.4 calories per gram versus sugar's 4. In the U.S. it is excluded from "Total Sugars" and "Added Sugars" on nutrition labels by FDA guidance. It is the backbone sweetener of every dessert on this site, and this page is the full, sourced case for that choice — including its limits.
Allulose is the answer to a question every other sweetener fails somewhere:what if a lower-calorie sugar kept useful bulk, browning and freezing behavior?
The chemistry, in plain language
Allulose is a monosaccharide — a single sugar unit, the same class as glucose and fructose — with the identical chemical formula as fructose (C₆H₁₂O₆). The difference is one flipped hydroxyl group at carbon 3, which makes allulose a C-3 epimer of fructose: a mirror-flip at a single position. That one flip is enough to break the key-in-lock fit with the human enzymes that shepherd fructose into energy metabolism, while leaving nearly every physicalproperty — solubility, browning chemistry, freezing-point depression, bulk — essentially sugar-like.
This is why the "is it artificial?" framing misses. Allulose isn't a lab-invented molecule like sucralose or aspartame; it's a naturally occurring sugar that was first identified decades ago and exists in small quantities in figs, raisins, jackfruit, wheat, maple syrup, and brown sugar. What's new is our ability to make it affordably at scale.
How it's made
Nature makes allulose in trace amounts, so extraction was never viable. The breakthrough came from rare-sugar researchers at Kagawa University in Japan — Ken Izumori's group — who identified an enzyme (D-tagatose 3-epimerase) that performs exactly that one carbon-3 flip. Commercial production runs that enzymatic conversion on fructose, usually from corn: starch → fructose → enzyme flip → allulose, then filtration and crystallization. It's the same production logic as making cheese with rennet or high-fructose syrup with isomerase: an enzyme rearranging a food molecule.
Calories and the label math
Human studies tracking labeled allulose found that most of an ingested dose is absorbed in the small intestine and then excreted unchanged in urine, with the unabsorbed remainder passing through with minimal fermentation — the molecule enters, sightsees, and leaves (Iida et al., Metabolism, 2010). FDA guidance recommends using0.4 kcal/g for labeling purposes and states that the agency intends to exercise enforcement discretion when manufacturers exclude allulose from "Total Sugars" and "Added Sugars" on U.S. Nutrition Facts labels. Allulose remains included inside "Total Carbohydrate."
This site reports a separate calculation: net carbs = total carbs − fiber − allulose. “Net carbs” is not an FDA-defined label declaration. Every nutrition panel prints allulose on its own line so readers can see the components and apply their preferred convention.
Blood sugar and insulin
Because human fructose-metabolizing enzymes don't accept allulose, it produceslittle to no rise in blood glucose or insulin in controlled human feeding studies. More interestingly, several small randomized trials report that a modest dose of allulose eatenalongside a carbohydrate load slightly blunts the post-meal glucose rise — a hypothesized effect on intestinal glucose handling and hepatic enzymes that meta-analyses of controlled feeding trials have found directionally consistent but modest in size. The honest summary: neutral-to-mildly-favorable for glycemic control, clearly better than sugar, and not a treatment for anything.
Safety and regulatory status
In the U.S., manufacturers have submitted allulose GRAS notices to FDA. FDA issued “no questions” response letters, beginning with GRAS Notice GRN 400 for D-psicose in 2012 and followed by additional notices from other producers. A no-questions letter is not an FDA approval; it means the agency had no questions at that time about the notifier's GRAS conclusion for the described uses.
What doesn't exist yet, and you should want named plainly: long-term human outcome trials. Allulose has short-term human safety and tolerance data, animal toxicology, and a metabolic argument for inertness — it does not have 20-year cardiovascular endpoint studies, because nothing this recent does. We use it anyway, in food-scale doses, with the grams printed on the label; that's a considered position, not a proof of immortality.
GI tolerance: the actual numbers
Every poorly absorbed carbohydrate — fiber, sugar alcohols, allulose — has the same failure mode at high doses: unabsorbed molecules pull water into the gut (osmotic effect), and the result is bloating, gas, or diarrhea. A small non-randomized study in 30 healthy young adults (Han et al.,Nutrients, 2018) proposed the following maximum single and total daily intakes. Treat them as study reference points, not guaranteed symptom-free limits for every person:
| Study reference | Amount | 60 kg / 132 lb | 75 kg / 165 lb | 90 kg / 198 lb |
|---|---|---|---|---|
| Single sitting | 0.4 g per kg bodyweight | 24 g | 30 g | 36 g |
| Across a day | 0.9 g per kg bodyweight | 54 g | 68 g | 81 g |
The twelve protein ice cream recipes provide about 17.5 to 32.5 g allulose per serving. That range can approach or exceed the study's 0.4 g/kg single-dose reference depending on body weight and individual tolerance; a whole pint doubles it. The labels print allulose grams so you can make that calculation. If you are unfamiliar with allulose, start below a full serving rather than assuming the study reference is a personal guarantee. A tablespoon of thecaramel sauce, for scale, contains about 9.8 g allulose.
EU and UK regulatory status
Allulose is not authorized on the EU Union list of novel foods. In its 2025 scientific opinion,EFSA concluded that the safety of D-allulose could not be established from the submitted dossier. That conclusion is not the same wording as a statutory ban, but it is not merely an administrative queue either. The ingredient cannot be marketed as an authorized novel food in the EU unless a future application resolves the evidence gaps and receives authorization. In the UK, application RP-1130 remains under risk assessment as of this update. Check the current regulator lists before buying, importing or formulating with allulose in either jurisdiction.
In the kitchen: what it actually does
| Property | Allulose behavior | What it means for you |
|---|---|---|
| Sweetness | Often described as roughly 70% of sucrose | Start around 1.3–1.4× by weight for similar sweetness, then adjust to taste; 1:1 blends vary by brand |
| Browning | Browns readily and can react faster than sucrose in some formulations | Watch color and temperature rather than applying one universal oven adjustment |
| Freezing | Contributes meaningful freezing-point depression as a small dissolved sugar | Final texture still depends on concentration, other solids, stabilizer and processing |
| Moisture | Hygroscopic and relatively resistant to crystallization | Can help cookies stay soft and sauces remain smooth; no formulation is immune to graining |
| Mouthfeel | Does not have erythritol's strong cooling effect | The finished flavor reads more like ordinary sweetness |
| Bulk | Provides gram-scale dissolved mass | Useful where high-intensity sweeteners cannot replace sugar's physical volume |
Two demonstrations of that table worth your time: the4-ingredient real caramel (browning + anti-crystallization + freezer-softness in one jar), and thetwelve-flavor ice cream library (freezing-point depression as dessert). For conversions in either direction — sugar recipes to allulose or back — use theconversion calculator.
Against the field, briefly
| Allulose | Erythritol | Stevia / monk fruit | Sugar | |
|---|---|---|---|---|
| kcal/g | ~0.4 | ~0.2 | ~0 | 4 |
| Sweetness vs sugar | 0.7× | 0.7× | 200–300× | 1× |
| Browns / caramelizes | Yes | No | No | Yes |
| Freezer behavior | Meaningful freezing-point effect | Can recrystallize and add graininess | Little effect at trace doses | Meaningful freezing-point effect |
| Cooling sensation | None | Strong | None | None |
| Class | Rare sugar | Sugar alcohol | High-intensity extract | Sugar |
On erythritol, one more honest note: a 2023 study in Nature Medicine (Witkowski et al.) reported an association between high circulating erythritol levels and cardiovascular events, plus platelet effects in lab work. Association isn't causation and erythritol is also made by your own body — but it moved many people to prefer alternatives, and it's part of why blends have been reformulating toward allulose. We use erythritol nowhere on this site; the freezer physics alone had already decided that.