Rice-koji amazake

A sweet drink without the sugar

Amazake is a pale, cloudy Japanese drink of cooked rice, and its defining quality is a sweetness that arrives without any sugar being added. Japan's Ministry of Agriculture, Forestry and Fisheries counts it among the country's traditional beverages and describes it as a fermented drink that is strongly sweet and rich in nutrients. The Japanese name, 甘酒 (amazake), carries the character for alcohol, which misleads readers in English; the ministry states plainly that because the drink holds no alcohol, or only trace amounts, it is not classified among alcoholic beverages at all. The name records what the drink descends from, not what is in the cup.

The rice version has its own name, because there are two ways to make something called amazake. The ministry names them to keep them apart: 麹甘酒 (kōji-amazake), built on rice inoculated with koji mold, and 酒粕甘酒 (sakekasu-amazake), built on the pressed cake left behind when sake is drawn off. They share a word and a place in the kitchen, and they reach that place by entirely different routes. Knowing which one a recipe means is the first thing a maker needs to settle.

The formula on this page is koji amazake, and it is deliberately minimal: cooked white rice, rice koji and water, held warm for most of a day. No sugar, no salt, no flavouring, no second culture. The sweetness is coaxed out of the rice itself by enzymes the koji carries. The home method followed here finishes with a mixture so loose it is described as a runny, rice-flecked porridge — thin enough to drink from a mug, thick enough that softened grains stay visible in it. It is a drink you can also eat with a spoon, and it is served hot.

Two clarifications belong at the start, because this drink is often described carelessly. The change that sweetens it is enzymatic rather than microbial: the British Columbia Centre for Disease Control's guideline on koji foods treats the warm stage as saccharification, the conversion of starch to glucose by amylase, and declines to call that stage fermentation at all. The ministry, meanwhile, calls the drink a fermented beverage. Both statements stand if the terms are kept apart — a mold grown on rice supplies the enzymes, and the enzymes do the sweetening. Whether living organisms are still active in the finished cup is a separate question, and the sources read here answer the first question but not that one.

The method recorded below is a modern domestic one, with cup-and-ounce quantities and a kitchen appliance holding the temperature. It is worth reading against the ministry's plainer account of the same drink, because the two agree on every essential and differ only in how precisely the numbers are written down.

Origins in the shadows of several theories

Nobody has settled where amazake began, and the careful sources say so directly. The ministry's account opens by noting that several origin theories exist, then offers the best-known candidate without endorsing it. The chronicle *Nihon Shoki*, compiled in the Nara period, describes a drink written 天舐酒 and read *ama no tamuzake*. The ministry's reading of that passage is that it was a sweet rice-koji drink of low alcohol, closer in kind to today's amazake than to modern sake. If that identification holds, the ancestor of this sweet drink sits on the koji-and-rice side of the family rather than the brewing side, and the line between those two was still soft when the chronicle was written.

That is the full extent of what the evidence supports, and it should be reported that way. The chronicle passage is presented by the source as one theory among several, not as a dated first appearance of amazake under its own name, and no source consulted for this page fixes a century at which the drink can be pinned down. The period field of this record is therefore left unestablished rather than filled with a guess. What the sources do settle is the drink's home ground: it is treated as indigenous to Japan, resting on rice as the staple crop and on koji as a foundational element of traditional Japanese cooking. Its identity is tied to two materials that define that kitchen.

The geography inside Japan is more legible than the deep history, and it traces the shape of the fermentation trades. Koji amazake, the ministry observes, is associated with regions where koji-based crafts such as miso and soy sauce are deeply rooted. Those were households and workshops that already grew mold on rice as routine work, so a sweet rice drink was a short step away. Sake-lees amazake clusters where breweries are numerous, because the lees are what remains after sake is pressed out of the mash. The two drinks distribute themselves along two economic lines: one follows the koji maker, the other follows the brewer and the residue the brewer is left with.

That division explains how each drink was historically obtained. The sake-lees version is essentially a way of using what the brewery has left over, and its recipe shows it — lees, water and sugar, with no conversion step needed because brewing has already done the work on the grain. Koji amazake is the more self-sufficient of the pair. It needs cooked rice and a supply of koji, and koji is the same culture a household kept for its miso paste. A family that made its own paste already owned everything the sweet drink required.

The ministry's Japanese-language account places widespread everyday drinking in the Edo period. Until roughly then, when sugar was not yet a common household staple, the drink mattered as a rare source of sweetness rather than as a treat. What it was drunk *for* has shifted more than how it is made.

The modern home recipe on this page cannot be traced back through any of that. It is a contemporary formulation, and nothing in the sources dates its particular ratio. The ministry's description of the same drink gives no quantities at all, only a process and an overnight hold. The recipe is presented here as a present-day adaptation of a long-standing practice, not as a recovered period formula.

Enzymes, not yeast: how rice turns sweet

Rice is mostly starch, and starch is not sweet. A grain of plain cooked rice tastes bland because its carbohydrate is locked in long chains the tongue cannot read as sugar. Amazake begins when those chains are cut short. The ministry attributes the drink's sweetness to enzymes produced by *Aspergillus oryzae* acting on starch, which is the bulk of the rice. That mold is koji, the culture called 麹 (kōji) in Japanese. It does not sweeten the drink by living in it; it sweetens it by having grown on rice beforehand and leaving a load of enzymes behind in the grain.

The BC CDC guideline explains where those enzymes come from in more detail. Rice koji is made by polishing brown rice, steaming it, cooling it, and only then introducing *A. oryzae* spores, which grow into the kernels. The temperature at which that growth is held shapes what the mold produces. For amazake, the guideline records incubation at 35 to 40°C, warmer than the 27 to 30°C used for miso koji, because the higher range favours amylase. The same mold on the same grain, kept a few degrees warmer, is steered toward the enzyme family this drink needs. Koji grown for a savoury paste and koji grown for a sweet drink are not the same tool; they are different outcomes of a temperature decision made days before the drink is mixed.

Once koji is stirred into cooked rice and water, what follows is chemistry carried out by those enzymes. The guideline describes the koji being diluted with water and held at 50 to 60°C, with amylase converting rice starch into glucose, and it names that stage saccharification rather than fermentation. The distinction matters. Saccharification is the cleaving of starch into sugars; it requires no living cell at the moment it happens, only the enzyme the mold supplied. This is also why the drink can be intensely sweet while holding little or no alcohol. No yeast has been asked to convert sugar into ethanol, and the recipe never introduces one.

The sugar that appears is not a single compound. The guideline reports at least twelve oligosaccharides among the products, grouped loosely under "glucose," and says their proportions shift across a range from 45 to 70°C. So the sweetness of amazake is a mixture — glucose alongside shorter and longer chain fragments — and the balance within that mixture is sensitive to how warm the mash was kept. Two makers working with the same rice and the same koji can end up with different sugar profiles purely because one held the pot a few degrees cooler. The sources establish that the variation exists and that temperature drives it; they offer no target profile and no way to measure one at home.

A mild acidity does develop on its own. The guideline notes citric and succinic acids among the products of the process, which accounts for the faint tang underneath the sweetness in a well-made cup. That tang is not strong enough to preserve the drink, a point taken up below. It is also worth separating the drink's chemistry from its texture: the grains in finished amazake are soft because they were cooked and then held warm for hours, not because enzymes dissolved them. The conversion acts on the starch inside the grain rather than on the grain's structure, which is why the finished drink still shows recognizable rice.

The recipe source describes adding the koji as the point where fermentation starts. Read against the ministry and the safety guideline, that is loose language for a precise event: the moment enzymes meet the starch they will act on. *Aspergillus oryzae* is a mold, and the recipe page's description of koji as a bacterial culture misstates what koji is; the ministry and the guideline both identify the organism by name. The enzymes do the work. The mold that made them stays upstream of the pot.

Three ingredients and what each one does

The recipe asks for half a cup of cooked white rice, four ounces of rice koji and one and a half cups of water. That is the entire list. No rice variety is named beyond "white," nothing salty appears, and no final volume is stated, so the batch stands as published rather than as a ratio a reader can safely double.

The rice is both substrate and flavour base. It supplies the starch the enzymes attack, and after hours of warm holding it supplies the soft grains suspended in the finished drink. The ministry frames the ingredient in the widest possible terms, as the Japanese staple food, and the sources consulted here say nothing more specific about grain type, soaking, or whether the rice is boiled or steamed before use. That silence is worth stating rather than filling: a reader who prefers a particular rice should know the evidence does not rank one against another for this drink.

The koji does the real work, and it does most of it before it ever reaches the pot. It is rice on which *Aspergillus oryzae* has been grown until the grains are loaded with amylase. The quantity here is strikingly generous — four ounces of koji against half a cup of cooked rice. No source gives a dosage rule or a koji-to-rice percentage for amazake, so this proportion is the recipe's own rather than a documented standard. What the guideline does establish is that koji meant for amazake is a specific thing, incubated warm enough to favour amylase, so a culture raised for a savoury paste is not the same tool in different packaging.

Koji is bought rather than kept. The recipe source notes that it turns up online and in Japanese grocery stores, and that it is usually sold dried. Drying suits a home maker, and it is why the method asks for the koji to be crumbled before going in: dried grains clump, and the enzymes act on contact, so breaking the clumps apart spreads the activity through the mash. The safety guideline treats koji as a supply-chain ingredient with real limits. It requires the *Aspergillus* to come from an approved commercial source, and it allows a kept starter to be refrigerated for roughly a month, or frozen, with some loss of activity acceptable as long as contamination is kept out. A packet that has sat in a cupboard far longer than that, or one carried forward from an earlier batch, is a different proposition.

Water is the medium. It loosens a dense mass of rice and koji into something that can be stirred and held evenly, and the guideline's own description of the stage treats dilution as part of the process rather than as an afterthought. Enzymes travel through liquid; a dry or pasty mash gives them less contact with the starch. The one and a half cups here is the only water figure in the sources, and no target consistency accompanies it, so the amount is the recipe's choice, not a specification.

Nothing sweetens this drink, and that is the whole point of the formula. The only version in the sources that takes added sugar is the sake-lees kind, which is built from lees, water and sugar — sugar goes in because there is no conversion step to generate it. In koji amazake the sugar is manufactured inside the pot, and every degree of holding temperature and every extra hour changes how much of it appears.

Why the hold is the whole recipe

Everything in this method happens during one long, warm hold, and the two published accounts of it agree more than they differ. The ministry's version is the spare one: cooked rice with water and koji stirred together, covered, and left overnight at 55 to 60°C, with a note that stirring partway through speeds the process. The home recipe gives the same shape with numbers attached — cover and hold at 130 to 150°F (54 to 66°C) for eight to twelve hours, stirring now and then and keeping the temperature inside that band the whole time. The guideline's band for the stage, 50 to 60°C, sits beneath both. Three sources, three slightly different ranges, and the overlap is obvious: a warm mash, kept warm for most of a day, never allowed to cool off and never allowed to get hot.

That band is the recipe's genuine difficulty, and it exists because enzymes have preferences. The conversion runs while the mash stays warm. Let it drift toward room temperature and the work slows to a crawl, which is why the ministry specifies a maintained overnight temperature rather than a pot left on the counter. Let it climb well past the band and the enzymes lose their grip on the starch. The guideline's observation that the sugar mixture varies across 45 to 70°C is the practical consequence: the holding temperature quietly decides what kinds of sugar end up in the cup, not only how fast they appear.

Stirring appears in both descriptions, and the ministry states its purpose outright — it promotes the process. In a mass of rice and koji, contact between enzyme and starch is everything. Turning the mash over redistributes grains that have settled, evens out cooler pockets near the wall of the vessel, and gives the enzymes fresh surface to work on. This is a physical operation rather than a biological one. Nothing is being aerated for the benefit of a living culture, because the stage is a chemical conversion.

Holding a pot between 54 and 66°C for eight to twelve hours is not something a stove manages unaided, which is why the recipe names its equipment: a thermal pot or a rice cooker, worked with a wooden spoon. A rice cooker on a warm setting, or a thermal pot that keeps its heat, performs the job the ministry's account leaves unnamed. The recipe asks that the temperature be maintained continuously, which in practice means checking it, yet the sources give no instrument specification and no endpoint test — no iodine test, no sweetness reading, no way to confirm the starch has gone. Time and temperature are the only guidance offered.

The end of the hold is marked by a deliberate rise in heat. The recipe instructs raising the mixture to 167°F (75°C) or higher, and gives the reason: stopping the conversion keeps the drink from turning sour. The ministry's account of koji amazake does not mention this step at all, so it belongs to the home method, but it is consistent with the chemistry the guideline lays out. Driving an enzyme-driven process to a finish means heating hard, and the source names its threshold.

What the pot should look like when it is done is described once, in the recipe source, as a very runny porridge flecked with rice. That is a useful picture, and it is also the only doneness signal in any of the sources. There is no target sweetness, no thickness standard, no clarification and no straining step. The drink passes from pot to mug or bowl exactly as it came out, grains and all. The ministry's description of the process stops at the same place, with no finishing operation of any kind.

One practical difference between the two accounts is worth naming for anyone who follows the home version. The ministry's hold is described as overnight, which is an open-ended span, while the home recipe bounds it at eight to twelve hours and then applies a stopping heat. The bounded version is the more repeatable of the two, and it is also the one that acknowledges that leaving the mash warm indefinitely invites something else to take over the sugar.

The other amazakes

Koji amazake is only one member of the family, and the differences between its relatives are sharper than the shared name suggests.

The most common alternative is *sakekasu amazake*, built on the pressed cake left behind after sake is drawn off. Its method could hardly be further from the one described above. Sake lees are dropped into boiling water, the heat is turned down and the mixture stirred until the lees dissolve, then the heat is raised and sugar added. No koji goes in, nothing is held warm for hours, and no conversion happens in the pot. The sweetness is bought with sugar, because the brewing that made the sake broke the grain down long ago. The lees carry a little alcohol, and the ministry notes that heating drives it off. The safety guideline treats that alcohol as a genuine variable rather than a culinary footnote: yeasts surviving from the sake-making can leave residual alcohol in this style of amazake, and the guideline says alcohol should be handled as a chemical hazard when the drink is planned commercially. Anyone serving this version to children, or to someone avoiding alcohol, has a real question to answer — and that question does not arise in the koji version at all.

A third type sits between the two in method and looks nothing like them in flavour. The guideline describes a sour amazake made with a koji starter built on lactic acid bacteria: it is prepared much as the sweet version is, then held at room temperature for anywhere between a day and a week, finishing below pH 4.6. Here a live microorganism is deliberately given the run of the pot, for far longer than the eight to twelve hours of the sweet version. The result is a drink soured by lactic acid rather than sweetened to the same pitch, and it reaches an acidity the sweet drink never manages. The difference comes from which starter was used and how long the culture was left alone, not from leaving the same pot resting a little longer.

That contrast is the most instructive one on the page. Sweet amazake and sour amazake can start from a similar mash and end on opposite sides of pH 4.6, and what separates them is the organism invited in and the length of the hold. It also means a maker cannot read safety off the finished glass. A mild, sweet-looking amazake is not the same kind of product as an acidified one, however alike they appear in the cup.

The ministry's regional note explains where each tradition took hold. Koji amazake belongs to the districts where miso and soy sauce culture runs deep; sake-lees amazake belongs to the sake-brewing country, following the byproduct. The ministry's Japanese-language account sharpens this: because lees differ from one brewery to the next, the character of the sake-lees drink varies with its source, and that variation is what gives the drink its regional identity. The ministry also lists separate regional beverages around Japan, including Kochi's *Hiyashi Ame*, which has its own entry rather than being a name for this one. Those pages are worth reading as neighbours of amazake, not as versions of it.

The catalogue holds no sake-lees recipe at present, so the variants above are described rather than linked. Nothing in the sources supports treating them as interchangeable: different ingredients, different processes, different alcohol and acidity profiles, and different reasons for being made.

Drinking it, and reading it safely

The serving instruction is short and literal: drink it hot, straight from the pot, in a mug or a bowl. Nothing is added at the end. There is no cold serving, no garnish, no flavouring and no straining. What comes out is thick and spoonable as well as drinkable, and its sweetness is the kind that wants something alongside it — the ministry's own description calls the drink strongly sweet, and the recipe source calls the finished mixture a very runny, rice-flecked porridge. A cup of it behaves less like a glass of juice than like a small, rich interlude. The sources consulted here name no specific dishes it accompanies, so no pairing is claimed; what they do establish is that the koji version has historically belonged to households already working with miso and soy sauce, and that is where it has been drunk. The ministry's Japanese-language account adds one habit worth knowing, because it runs against the modern expectation: today amazake is often thought of as a winter and New Year drink, but widespread everyday drinking is placed in the Edo period, when it was popular as a summer refreshment sold by street vendors.

Storage is where this record has to be blunt, because the drink's own chemistry works against it. Sweet amazake is full of free sugar, and the safety guideline's position is that the mild acids which form during the warm hold do not make it self-preserving: the pH can remain above 4.6, and for that reason the product has to be refrigerated. A pH above 4.6 means this is not a pickle, not a yoghurt, not a product whose acidity protects it on a shelf. The recipe's own advice — keep it sealed and cold for up to a week — is a home cook's figure rather than a validated shelf life, and the guideline that insists on refrigeration offers no storage duration for amazake at all. The honest position is that the drink must be refrigerated and that the length of time it keeps has not been established by any source read here.

The warm hold deserves the same clear-eyed look. The working band in this recipe, 54 to 66°C, is the range every source gives for the conversion; the ministry's is 55 to 60°C and the guideline's is 50 to 60°C. It is also a temperature at which a sugar-rich, near-neutral food sits for eight to twelve hours, and no source consulted here turns that into a validated control. The guideline's reasoning about rice points the same way: it warns that heat-stable toxins of the kind associated with rice, such as the cereulide produced by *Bacillus cereus*, survive boiling, which is why it insists that any soak running past four hours happen cold or acidified. A pot of rice that has stood in water overnight on the counter can carry a toxin that no amount of later reheating removes. That is a real consideration rather than a reason to call the method wrong, and it is worth carrying into the kitchen alongside the temperature band.

Two upstream points from the guideline bear directly on what a home maker buys and keeps. First, koji should come from an approved commercial source rather than being carried forward from an earlier batch. The guideline advises against using one batch of koji to raise the next, because doing so amplifies the risk of cultivating wild *Aspergillus* strains capable of producing aflatoxins; it also allows a kept starter to be refrigerated for roughly a month, or frozen, accepting some loss of activity provided contamination is kept out. Second, the soaking rule just mentioned applies to the rice before it becomes koji as well as to rice cooked at home. Both points are about ingredients, and both are cheaper to respect than to undo.

None of this can be judged by taste. A sour note in finished amazake tells the maker that the conversion was left running too long, which is exactly what the recipe's final heating step exists to prevent; it says nothing about whether the drink is safe. Sweetness, aroma, bubbles and appearance are qualities of the food, not measurements of it. The narrow summary the sources support is usable: make it hot, hold it warm, stop it hot, refrigerate it promptly, and do not assume a longer keeping time than the recipe claims without evidence behind it.

A final note on what this page does not claim. The ministry's description of the process is a summary drawn from an encyclopedia entry, so the 55 to 60°C figure it gives is a secondary report rather than a laboratory measurement. The home recipe supplies quantities but no yield. The guideline supplies safety reasoning but no amazake quantities, and its hazard table and critical control points are written for miso rather than for this drink. Between them the three accounts agree on the shape of the method and leave its exact numbers open, and that is how the record above is written.