Every twaróg recipe tells you to heat the soured milk gently and not let it boil. Almost none of them tells you what gently means, and the ones that do give a number do not agree with each other.
So I read twelve of them, Polish and English, and set the numbers side by side. The lowest says 36 C. The highest says 82 C. That is a 46 degree spread on the one step that decides whether you get a soft, moist cheese or a dry, squeaky one, and half of them give no number at all.
Not one of the twelve mentions that the disagreement exists. Each states its own figure as though it were the only one.
The twelve recipes, side by side
| Recipe | Temperature given | What it says |
|---|---|---|
| Polonist | 36 to 39 C | Stirring, 30 minutes to an hour |
| Ania Gotuje | 38 C | 30 to 60 minutes, or 40 C in the oven. Do not boil |
| Holly Trail | 38 C | Medium-low, 45 minutes to an hour, curd should cut cleanly |
| PrzyslijPrzepis | 50 C | In pulses with five minute rests. Slower gives a more delicate cheese |
| Polish Feast | 50 to 60 C | Do not stir at first |
| Moje Wypieki | 70 C | Do not let it boil |
| Polish Housewife | 82 C | Milk heated first, acid added off the heat |
| Kielbasa Stories | Near boiling | Heat until “little bubbles”, then add vinegar |
| Kulinarne Rozkosze | No number | Very low heat, do not boil, until curds form |
| Polish Your Kitchen | No number | Lowest setting on the hob, about an hour |
| Beszamel | No number | Fresh milk boiled, then poured into the soured milk |
| Wszechjedąca | No heating at all | Curds form from souring alone, drained cold |
This is two different techniques wearing one name
Five recipes sit between 36 and 60 C. Three sit at 70 C and above. One does not heat the milk at all and still makes cheese.
The split is not random. The gentle camp is mostly the Polish sources and the ones that use a live starter, kefir, soured milk or cultured sour cream. The hot camp is mostly the English language recipes built on buttermilk or vinegar, where the acid is added to hot milk rather than grown in it over two days.
One ferments the milk and then warms it just enough to gather the curd. The other heats the milk and crashes it with acid. They are aiming at different cheeses, and none of them says so.
What the temperature is actually doing
An acid-set curd is not the same thing as a rennet-set one. The acid has already dissolved out the calcium that held the protein network together, so what you have is loose clumps of casein with water trapped between them.
When you heat that, the first thing driven off is water, not protein. Push it further and the water boils away, the protein dries out and concentrates further, and that is exactly why firm paneer and queso blanco can be simmered or fried like meat without melting.
That is the whole trade in one sentence. Gentle heat barely starts the process and leaves a soft, moist, spreadable curd. More heat pushes water out and gives a firmer, drier one. Too much and you get the rubbery result people complain about, because rubbery is simply what this cheese becomes if you keep concentrating it.
And cheesemaking has a number for it. When the trade wants whey out of a curd it cooks the cut curd in its own whey at 55 C, the temperature that expels it. That is rennet curd for hard cheese rather than the acid curd we are making here, so it is not a line drawn for twaróg. It is still striking that every recipe in the gentle camp sits below 55 and every one in the hot camp sits above it. The two groups are not a gradient. They are on opposite sides of the temperature the trade uses deliberately when it wants a drier curd.
One more thing worth being clear about, because “drier” and “less creamy” sound like two faults and they are one. The fat does not go anywhere. It is the water that leaves, so the same fat ends up held in a stiffer, more concentrated curd. That is why a hot-method twaróg can be perfectly good cheese and still be the wrong cheese for a filling.
So neither camp is wrong. They want different things.
The 78 degree line, and why these recipes cannot settle it
There is a second thing that happens, and it has a precise temperature attached to it.
Milk carries a second protein besides casein, called lactoglobulin, and normally it stays dissolved and runs off with the whey. It unfolds at 78 C. Once unfolded it stops staying out of the way and attaches itself to the casein instead.
Follow that through and it predicts something you could test: a recipe heated past 78 C should keep protein that a recipe heated to 38 C throws away, so the hot camp should get more cheese out of the same milk.
These twelve do not show that, and I am not going to pretend they do. The 82 C recipe reports about 15 per cent of the milk’s weight as cheese, one of the lower figures here. The highest yield in the group, about 26 per cent, comes from a recipe that heats to 38 C. The lowest, about 11 per cent, comes from one of the gentlest.
The reason is that the yields are not comparable. Different milks at different fat levels, different amounts of starter, and draining times running from ten minutes to overnight under a weight. Draining swamps everything else, and you cannot see a protein effect through it.
So the 78 degree line is real chemistry that is not visible in published home recipes. Settling it would take one batch of milk split in two, heated to different temperatures, drained identically and both weighed. That is worth doing and I have not done it yet.
What I use
50 C, and I take the pan off at about 48, because the pan is hot and carries it the rest of the way on its own.
That is my sweet spot and I arrived at it by making the cheese, not by reading about it. I need this one for cooking: pierogi ruskie, sernik, a sweetened naleśniki filling. All three want it soft and creamy, and hotter makes it drier and less creamy, which is the one thing none of those three dishes forgives. My kitchen and the chemistry happen to agree, and the chemistry came second.
The number is not guesswork either. It is the figure in my own written recipe from 2020, and the photograph at the top of this page is my probe in the pan reading 47.9 C, just before I take it off the heat.
If you want the sliceable kind instead, the honest lever is not more heat, it is longer draining under a weight, which gets you there without the rubbery risk.
And if you have no thermometer, the old test beats any number: the pan should stay cool enough that you can hold it with both palms the whole time.
The full method is in my twaróg recipe.
How I did this
Twelve recipes with usable instructions, read in full in September 2026: Ania Gotuje, Beszamel, Wszechjedąca, Kulinarne Rozkosze, PrzyslijPrzepis, Moje Wypieki, Polish Your Kitchen, Polonist, Kielbasa Stories, Holly Trail, Polish Housewife and Polish Feast.
Three more could not be read and are left out rather than guessed at: pracowniaserow.pl returned 410 Gone, momsdish.com returned 403, and I could not find a twaróg page on kwestiasmaku.com.
Temperatures are quoted exactly as each source gives them. The yield percentages are worked out from the stated milk volume and finished weight rather than published as percentages, so treat them as indicative.
You might also like: