The Short Version First
You don't actually make yeast from scratch. You propagate Aspergillus oryzae spores into a solid grain culture, then optionally multiply that into liquid starter. The whole thing hinges on temperature control and cleanliness, nothing else. Most people fail because they rush the humidity step. I have made this culture for about a decade across different batches and setups. The basic process uses short-grain rice, water, and a spore source. I started with commercial koji koji blocks but switched to spore suspension after my third batch got contaminated with Penicillium. That taught me how fragile the early growth phase actually is. Here is what you need.
Short-grain Japanese rice or regular short-grain rice works fine. You need about 500 grams for a small batch. Filtered water, 600 milliliters. A spore source, either a commercial koji starter like Hongo no Ikari or spores scraped from a healthy koji block. Stainless steel or glass containers, never plastic if you can avoid it. A way to maintain 85 to 90 percent relative humidity during incubation. Something like a sealed container with a damp paper towel at the bottom. An incubation space holding 28 to 30 degrees Celsius. A thermometer. A spray bottle for misting. Rinse the rice thoroughly until the water runs mostly clear. That removes excess surface starch which otherwise creates a slimy texture that traps unwanted bacteria. Soak the rice for four to six hours. Drain completely. Steam the rice rather than boil it. You want the grains soft but individually separate, not mushy. That usually takes about 40 minutes over boiling water in a steamer insert. If you use a rice cooker, add only 60 percent of the normal water amount and use the quick cook setting if available. Spread the steamed rice on a clean stainless steel tray to cool. You are waiting for the temperature to drop to 30 to 32 degrees Celsius before inoculation. Put it any warmer and you kill the spores. Any cooler and the colony growth stalls out and contaminants move in first. I learned this the hard way when I tried to speed things up by spreading the rice on a warm counter. The resulting koji had a sour off-note from bacterial dominance because the spores never established quickly enough.
Inoculate at that 30 to 32 degree window. If you are using powdered starter, dust it evenly over the surface at about one gram per 100 grams of rice. If you are using a spore suspension, mix one teaspoon of spores into 100 milliliters of sterile cooled boiled water and mist that over the rice. Toss the rice gently with a sterile spatula so the spores distribute through the top two centimeters where oxygen is available. Koji is an aerobic organism. Bury it and it dies. Transfer the inoculated rice into your incubation container. Line the bottom with a damp but not dripping paper towel. Cover the rice loosely with breathable cloth or a lid that is not sealed tight. You need gas exchange. Place the container in your warm spot and check the temperature every few hours. Maintain 28 to 30 degrees Celsius. Mist the inside walls of the container if the humidity drops below 80 percent. The rice itself should stay moist but never sitting in free water. First sign of growth shows at about 24 hours as white mycelial threads covering the grains. By 36 to 48 hours the surface should be densely colonized with a fresh white to slightly cream colored fuzz. That is the mycelium and conidiophores developing. If you see green, blue, or black spots anywhere, discard the batch immediately. That is contamination and there is no salvaging it. I once had a batch develop a thin green ring around the edge of the container. Turned out the lid was slightly warped and allowed airborne spores to settle there. The rest of the batch was fine but I threw the whole thing out rather than risk it.
Get the Full Details

At 48 to 72 hours the culture is ready. The grains should be fully coated in white mycelium and emitting a sweet earthy smell, similar to fresh cut grass or steamed buns. Any ammonia smell means it is going bad. The produced enzymes, mainly amylase and protease, are now active inside the colony. That is what you need for saccharification when you actually make the wine. If you want to preserve the culture, dry it. Spread the colonized rice on a dehydrator tray at 40 degrees Celsius until the moisture content drops below ten percent. That usually takes about 12 hours. Store the dried koji in an airtight container in the freezer. It stays viable for about six months. At room temperature it lasts maybe three weeks before enzyme activity declines significantly. There are real limitations to this method. The humidity requirement is the hardest part to manage without dedicated equipment. A simple sealed box with a wet towel works but you have to check it twice daily. Temperature fluctuation during incubation causes uneven growth. You will get some colonies that are fully sporulated next to patches that are still just mycelium, and that inconsistency affects enzyme output. The other issue is spore viability. Cheap commercial starters sometimes arrive with low viability, especially if they have been sitting on a shelf in a warm environment. Always buy from a supplier who turns stock over quickly. I once used a starter that was six months past its printed date and the colonization took twice as long with weaker enzyme activity. The resulting sake had barely any sugar conversion and tasted thin and watery.
For home use the solid grain method is the most practical approach. Liquid starter propagation exists but it requires sterile technique that most people do not have the equipment or patience for, and the shelf life is measured in days rather than months. If you want consistency without managing humidity chambers, buying a quality commercial koji starter and propagating it through one or two home generations is reasonable. The enzyme profiles change slightly with each propagation cycle, but for most home brewing purposes the difference is negligible. The key things that separate a good batch from a failed one are temperature stability during the first 24 hours, cleanliness of all surfaces, and not adding more water than the rice can hold after steaming. Get those three right and the rest follows naturally.