Kojiology

Physics, thermodynamics & enzyme biology

Incubator Simulator

Koji is rice, barley, or soybeans inoculated with Aspergillus mould — the living enzyme factory behind miso, shoyu, sake, and shio koji. Set your chamber and batch — get safe bed thickness, the vent count or door gap that keeps it cool, and a 3-phase heat timeline with the actions each phase needs.

New to this? Start here

A koji run is about two days: steam the grain, cool it, mix in spores, then hold it warm and humid while the mould colonises it. The part that surprises people is that heating the box is only the first day's problem. Once the mould is growing it produces its own heat — a few kilos of koji is a small furnace — and from roughly hour 24 the job flips to getting that heat out before the middle of the bed passes 45°C and cooks the culture you just grew. That is why this tool is mostly about depth, vents and airflow rather than about a thermostat setting.

Ambient vs. core —
the air in the chamber vs. the middle of the grain. The core runs hotter, and it is the one that kills a batch — a probe thermometer in the bed is not optional.
Bed depth —
how thickly the grain lies. Deep beds trap their own heat in the middle, so there is a maximum depth for any given chamber.
Vents / air changes —
how heat, CO2 and moisture leave. Cool air enters low and warm air leaves high, so vents are needed at both levels for the flow to exist at all.
Mixing (te-ire) —
breaking the bed up part-way through. It sheds heat and water and separates grains that have started to mat together.
°C
m
kg
sets harvest deadline
Box geometry
m
m
m
m
Chamber · isometric
Set ambient this phase25°C · 70% RH · 131 air changes/hr

Needs 823 cm² of opening — 34% of the footprint, far too many holes to drill. The box stays sealed: thin the bed or split the batch instead.

Cool air is drawn in low and leaves high — vents have to exist at both levels or the stack effect has nothing to drive it. The dashed plane marks the safe bed depth.

Hydrate to
1.47 kg
+0.47 L water
Tane-koji dose
10 g
10 g/kg tane-koji · pure koji-kin is 1 g/kg
Bed depth
0.9 cm
max 5.2 cm
Predicted core
25°C
+0.3°C over 25°C ambient · 45°C red line — the mould denatures above it
Harvest by
48 h
cycle runs 48 h; mix at 24 h and 36 h
Expected yield
1.32 kg
after ~10% cycle loss
That yield figure assumes the roughly 10% of its mass koji loses over a cycle to respiration and evaporation. Weigh the batch when you pull it and write the number down: coming in near that loss is the cheapest signal you have that the batch actually worked.
The dose above is tane-koji, the common home/retail starter. Pure koji-kin is a tenth of it (1 g/kg, or 10 g/kg once diluted 1:10 into rice flour), and a saved tomo-koji batch needs more than either. They are not interchangeable gram-for-gram — see the Spore Inoculation tool.
Amylase strategy: higher moisture (40%), a thicker bed, and a warm start (38°C) — the koji grows aggressively (so-haze, mould coating the grains evenly) and pushes starch toward sugar, which is what makes amazake and sweet miso sweet. A thick bed holds its own heat, so watch the core, not the air.
Plan for an active water source at the peak.Cooling this batch at the heat peak takes about 131 air changes per hour in a 0.096 m³ chamber. That is normal for a passive box and it is not a fault in your setup — but it does mean dry room air is moving through fast enough to strip the humidity, so plan on an active water reservoir or a damp cloth over the bed rather than vent holes alone. A smaller batch or a cooler room brings the number down fastest.
Drilling is not the mechanism at this load.At the heat peak this batch needs about 34% of the footprint in vent area, which is far more hole than is practical to drill — and micropore tape, which every vent hole should have, cuts the airflow each hole delivers to roughly a sixth. That is true of a taped box at almost any batch size, so it is a property of the method rather than of your numbers. Plan to lift the lid at the peak under a damp cloth, or bring the heat load down: a thinner bed, or the batch split across more containers.
Want protease vs. amylase, not just the schedule?

Enzyme Balance is its own tool: strain, substrate, temperature, moisture, and bed depth each pull your koji toward sweetness or umami — see which lever is actually doing the work.

Open →

Dynamic operation schedule

Each phase's heat load and the vent action it needs.

0–18h2.9 W
Phase 1: Dormancy & germination

Spores wake up; almost no heat of their own yet, so this is the phase that needs external warmth. Keep humidity high so the surface never dries. First visible whitening lands around hour 18.

Set ambient38°C · 92% RH
9 open vent holes (7 cm²)
18–36h28.6 W
Phase 2: Hyperactivity (heat peak)

Metabolic burst — the koji now heats itself, so the job flips from warming to cooling. First mix (tefukumi) at hour 24: it breaks up the bed, sheds core heat, and refreshes oxygen. Drop the ambient and the humidity now.

Set ambient25°C · 70% RH
823 cm² open area needed — too many holes to drill (34% of footprint); thin the bed or split the batch
36–48h11.7 W
Phase 3: Maturation (synthesis)

Heat output declines and enzyme synthesis takes over — the lower humidity here is deliberate, it pushes the mycelium to produce enzymes rather than just grow. Second mix around hour 36 if the bed is dense. Harvest the moment a green/yellow tinge appears.

Set ambient25°C · 70% RH
337 cm² open area needed — too many holes to drill (14% of footprint); thin the bed or split the batch

The phases above say what the bed is doing; they do not say what it should look like while it does it. Haze Navigator owns that half — the six named growth patterns, what each one means for enzyme yield, and which lever moves you between them.

A probe thermometer in the substrate core and a hygrometer are non-negotiable regardless of setup — "blind" incubation with no measurement is the single most common cause of contamination and baka-hazé (soft, over-penetrated koji). This model is a physics guide, not a substitute for reading your own thermometer.