Vegan Gelato Without a Machine (Coconut, Actually Creamy)
Table of contents
Most vegan no-churn recipes fail twice over: once for being vegan and once for having no machine. Both failures have the same cause, which is missing solids, and both have the same fix. Here is a coconut base built to survive a domestic freezer.
What milk protein was doing, and what replaces it
A dairy gelato gets its texture from four contributors: milk fat, lactose, milk protein and added sugar. Take the dairy out and you lose three of them at once. Fat is easy to replace, and coconut replaces it generously. The other two are the reason so many vegan batches come out icy.
Milk protein binds water at the molecular level and stabilises the air-serum interface. Nothing in the plant kingdom does both jobs as cheaply. What works instead is a division of labour across three ingredients, and a no-machine recipe needs all three.
Bulking solids that add no sweetness. Maltodextrin DE19 has a PAC of 0 and a POD of about 10, so it contributes solids and body without touching either the freezing point or the sweetness. In a base that has no lactose, that role has to be filled deliberately. The full profile is in maltodextrin in gelato.
Soluble fibre for mouthfeel. Inulin is a chicory-root fibre that behaves in a frozen matrix a little like fat does: it builds a fine particle network that reads on the palate as creaminess. It carries PAC 0 and POD 0, so like maltodextrin it buys structure for free.
Gums, at a higher dose than usual. Locust bean and guar raise the viscosity of the unfrozen phase and physically slow crystal growth. In a machine they are a refinement. Here they are load-bearing, and the ratio matters, which is why the vegan stabiliser blend is worth building rather than buying.
Quick reference. No milk protein means the solids have to come from somewhere else. Target 41% total solids, 10% fat, PAC near 23 and POD near 17, and expect to use maltodextrin and inulin to get there.
The base, per 1000 g
| Ingredient | Weight | Why it is there |
|---|---|---|
| Full-fat canned coconut milk, 20% fat | 520 g | 104 g of fat plus flavour |
| Water | 181 g | Brings the mix to weight |
| Sucrose | 118 g | Sweetness and freezing point |
| Dextrose anhydrous | 60 g | Anti-freezing power at low sweetness cost |
| Maltodextrin DE19 | 70 g | Solids and body, no sweetness, no PAC |
| Inulin | 45 g | Fibre standing in for milk protein |
| Locust bean and guar blend | 5 g | Crystal control |
| Fine salt | 1 g | Flavour lift |
That lands at roughly 41.3% total solids, 10.4% fat, 17.8% sugars, a PAC near 23 and a POD near 17 per 100 g of mix. Total solids that high are not a typo. A no-machine base needs them, for the reason set out in the total solids guide: every point of solids you add is a point of water that never has to become ice.
The 60 g of dextrose is the single most important line in the table. It does the job described in sucrose versus dextrose, buying freezing-point depression at three-quarters the sweetness cost, and a vegan base needs more of it than a dairy one because there is no lactose contributing PAC in the background. Drop it below about 45 g and the gelato comes out of a domestic freezer as a solid block.
The gum dose, 0.5% of the mix, is roughly double what a machine-churned gelato needs. That is deliberate and it is the same dose the pistachio no-machine base uses.
Choosing the coconut milk
Not all cans are the same ingredient, and the label tells you which one you have bought.
Look for full-fat coconut milk with no added water beyond what the extraction needs, ideally listing coconut extract and water only. Canned full-fat coconut milk runs around 20 to 24% fat and roughly 2 g of protein per 100 g according to USDA FoodData Central, which is why the earlier point about protein is not a rhetorical flourish. There is almost none. The ingredient profile is broken down in coconut milk for vegan gelato.
Avoid the cartons sold as coconut drink. Those sit nearer 2% fat, and substituting one for the other drops the recipe from 10.4% fat to under 2%, which no amount of gum will rescue.
Guar gum on the ingredient list of the can is fine and common. It contributes to the total gum load, so if your can lists it, take the blend down from 5 g to 4 g.
One flavour note. Coconut fat is about 90% saturated and dominated by lauric acid, and refined coconut fat melts at roughly 24 °C. That is well below body temperature, so it releases cleanly, but it also means the fat is fully solid at freezer temperature and contributes hardness rather than softness. If you want the coconut flavour muted rather than forward, use a refined can rather than a virgin one; virgin coconut milk carries far more of the aromatic short-chain compounds.
Making the mix
Blend the sugars, maltodextrin, inulin, gums and salt together dry. This step is not optional. Gums and inulin both lump on contact with liquid, and inulin in particular hydrates into stubborn white beads that survive every amount of whisking after the fact.
Whisk the dry blend into the water while it is cold, then add the coconut milk and heat, stirring, to 82 °C. Hold for two minutes. There is no dairy in this mix, so the pasteurisation equivalences that 21 CFR 135.3 sets for frozen dessert mixes are not the governing constraint. The heat is there to hydrate the gums fully and to dissolve the inulin, which needs above about 70 °C to go into solution cleanly.
Do not boil it. Coconut milk emulsions break at a rolling boil, and once the fat separates it will not come back with a whisk. If you see oil beading on the surface, take it off the heat immediately and blend it with a stick blender for 30 seconds while it is still hot.
Then chill hard. Ice bath, down to 4 °C inside 90 minutes, and age at 4 °C for at least four hours and ideally overnight. Ageing is where the gums finish hydrating and the coconut fat crystallises, and it is the cheapest quality improvement available to a home batch. Skipping it costs more texture in a vegan base than in a dairy one, because there is no protein to compensate.
The aged mix will be thick and may look separated at the top. Whisk or blend it smooth before freezing.
Freezing it without a machine
A batch freezer drops the temperature fast and shears the mix while that happens. At home you can only do those separately, which is why the recipe had to change before the method could. The two routes below are the same ones that work for the chocolate water base, and both work here.
| Shallow pan | Freeze and blitz | |
|---|---|---|
| Attention needed | Every 30 min for 3 h | Two short sessions |
| Total elapsed time | About 5 h | About 7 h |
| Texture | Good, slightly coarse | Closest to a machine |
| Fails when | An interval is missed | The fat is over-worked |
Method A: shallow pan and a fork
Pour the aged mix into a wide metal pan no deeper than three centimetres and put it in the coldest part of the freezer. Metal pulls heat out several times faster than plastic or glass, and depth is the biggest single lever you have on crystal size.
Whisk hard every 30 minutes for the first three hours, reaching into the corners where the mix freezes first. Six interventions is the working minimum. When the texture goes from slush to a stiff, spoonable paste, scrape it into a chilled container, press cling film onto the surface and harden for two hours.
Method B: freeze solid, then blitz
Pour the aged mix into ice cube trays and freeze solid, four to six hours. Tip the cubes into a food processor and run it. The mixture goes through a gravelly stage, then a paste stage, then turns smooth and glossy.
Coconut has a specific failure mode here that dairy does not. Run the processor too long and the friction warms the mix past 24 °C at the blade, melting the coconut fat out of its crystals, and it re-sets as grainy specks. Stop at the glossy stage. If you can smell coconut oil rather than coconut, you have already gone too far. The two methods are compared against a domestic machine in no-churn gelato methods compared.
Faults and fixes
Smoothness in a frozen dessert is a question of ice crystal size. Crystals above roughly 50 micrometres register on the palate as iciness, while a well-made product averages nearer 20 to 25, a point Marshall, Goff and Hartel make throughout Ice Cream. Every fault below traces back to that, and the diagnostic path through them is why is my gelato icy.
| Symptom | Most likely cause | Fix in the next batch |
|---|---|---|
| Icy and coarse | Solids too low, or coconut drink used | Check the can is 20%+ fat |
| Rock hard from the freezer | PAC too low | Add 10 g dextrose, take 10 g off sucrose |
| Soupy, will not set | Too much dextrose | Cut dextrose to 45 g |
| Grainy, small hard specks | Coconut fat melted and re-set | Stop the processor sooner |
| Oily film on the surface | Mix boiled and the emulsion broke | Blend hot, keep below 85 °C |
| Chalky or gritty on the tongue | Inulin not dissolved | Blend it dry, heat above 70 °C |
| Gritty from day four | Heat shock in the freezer | See below |
That last row is not a recipe fault. A domestic freezer cycles through a defrost phase, and every cycle partly melts and refreezes the surface of the tub. That is heat shock, and it turns a good batch gritty by day four when nothing was wrong on day one. Store the tub at the back of the freezer rather than in the door, keep it full, and eat a no-machine batch within three days.
If you want to push further, the aquafaba route adds a genuine protein-like foaming agent to the same base and is covered in aquafaba in vegan gelato. Other plant bases, and the bilanciamenti that go with them, are in the complete vegan gelato guide.
