Silken tofu contains 5.0 g of protein per 100 g according to the Swiss Food Composition Database. Firm tofu contains 14.7 g. That is a threefold gap, and it changes what you can honestly say about a silken tofu dessert: it replaces cream beautifully, it does not replace a protein source.

What it genuinely replaces is fat. In a cheesecake, 500 g of silken tofu brings 265 kcal and 2.0 g of saturated fat, where 500 g of full-fat cream cheese brings 1,755 and 96. The saving is enormous and entirely real. It costs you protein, which falls from 55.5 g to 25 g across the same quantity.

The rest of this page weighs the six standard recipes instead of calling them light.

What silken tofu is, and how it differs from firm tofu

Silken tofu is soy milk curd that was never pressed. Every tofu starts the same way: soaked soybeans, ground, heated, and the resulting soy milk coagulated with a calcium or magnesium salt. What happens next splits the two families. Firm tofu is moulded and pressed, which drives out the whey and concentrates the dry matter. Silken tofu sets directly in its package, with no draining at all.

The composition shows the consequence, per 100 g:

	Silken tofu	Firm tofu
Energy	53 kcal	148 kcal
Protein	5.0 g	14.7 g
Fat	2.9 g	8.5 g
Saturated fat	0.4 g	1.3 g
Calcium	49 mg	120 mg
Water	87.6 g	75.1 g

Silken tofu is 87.6% water. That is precisely what makes it useful in baking: blended, it delivers a creamy texture without contributing fat. It is also what rules out describing it as a protein food.

One useful detail on calcium: the content depends on the coagulant. Firm tofu made with calcium sulfate shows 220 mg per 100 g, against 72 mg with nigari, which is a magnesium salt. The same gap applies to silken tofu depending on the process, and the front of the package rarely mentions it. The ingredient list does: calcium sulfate or magnesium chloride.

What the six desserts actually weigh in at

The recipes below are the ones that appear everywhere for silken tofu. Calculated on Swiss Food Composition Database values, they come out like this per portion:

Dessert	Portion	kcal	Protein	Sat fat	Free sugars
Raspberry fruit cream	174 g	115	5.8 g	0.5 g	7.7 g
Mango fruit cream	174 g	130	5.4 g	0.5 g	7.7 g
Vanilla flan	196 g	148	7.6 g	0.6 g	16.0 g
Cocoa flan	200 g	162	8.4 g	1.1 g	16.0 g
Chocolate cream	135 g	214	6.4 g	5.3 g	18.4 g
Cheesecake	141 g	322	7.8 g	9.8 g	21.8 g

A factor of three separates the raspberry cream from the cheesecake, and all of them carry the same “light silken tofu dessert” label. The difference has nothing to do with the tofu, which is near identical across the recipes. It comes from what gets added: 100 g of dark chocolate in the chocolate cream, a biscuit base bound with coconut oil in the cheesecake.

Free sugars are the real subject, not calories

The NHS defines free sugars as all added sugars plus those in honey, syrups and fruit juices. Maple syrup and agave syrup are named explicitly, on exactly the same footing as white sugar. Only the sugars naturally present in whole fruit, vegetables and milk are excluded.

That single definition dismantles the most common claim made about these desserts. Swapping sugar for agave syrup does not change the category, only the flavour and, marginally, the sweetening power.

The recommended limit is 30 g of free sugars a day for an adult, 24 g from age 7 to 10, 19 g from 4 to 6 and 14 g from 2 to 3. No added sugar at all below one year.

Measured against that, a slice of cheesecake at 21.8 g of free sugars accounts for 73% of an adult’s daily limit, and on its own exceeds the limit for a child aged 4 to 6. The flan at 16 g takes up slightly more than half. The raspberry cream, at 7.7 g, takes a quarter.

The practical conclusion fits in one line: in these recipes the variable you control most is the sweetener, not the tofu. Halving the syrup in a fruit cream takes it from 7.7 to 3.9 g of free sugars, and the fruit more than covers the difference in taste.

What the swap genuinely changes

The real case for silken tofu in baking is best made by comparing it with what it displaces, per 100 g:

	kcal	Fat	Of which saturated	Protein
Silken tofu	53	2.9 g	0.4 g	5.0 g
Fortified soy yoghurt alternative	40	2.1 g	0.3 g	3.9 g
Whole cream	334	34.8 g	20.9 g	2.0 g
Full-fat cream cheese	351	32.1 g	19.2 g	11.1 g
Double cream	468	50.0 g	30.0 g	1.9 g

Against whole cream, silken tofu brings six times fewer calories and roughly fifty times less saturated fat, while supplying two and a half times more protein. That is a decisively good swap, and it needs no exaggeration to be worth making.

Against cream cheese the picture is more mixed: energy and saturated fat collapse, but protein drops from 11.1 to 5.0 g. In a cheesecake the swap is still clearly favourable. In a dish where the cream cheese was carrying the meal’s protein, you need to account for that elsewhere.

A word on calcium, often used as an argument: silken tofu supplies 49 mg per 100 g, which is less than a fortified soy yoghurt alternative at 120 mg. If calcium is the goal, fortification matters more than the choice of tofu. Our complete guide to plant proteins covers sources and reference portions.

The texture does not come from the tofu

The notion that silken tofu is “naturally gelled” comes up constantly. At 87.6 g of water per 100 g, it is mostly naturally liquid once blended. What holds a flan together is the agar.

Agar is a gelling agent extracted from red seaweed. It has to reach a rolling boil to hydrate, and it only sets on cooling, at around 40 °C. That is why the flan recipe boils the plant milk with the powder for a full minute before blending it with the tofu: heating the finished mixture afterwards does not work better, and heating it less does not work at all.

The creams contain no agar. They hold because melted chocolate resolidifies when cold, or simply because fruit purée is viscous. This is why a fruit cream without chocolate stays soft and will not unmould: that is normal, not a failure. Our page on agar-agar and gelling without gelatin gives the doses for each level of firmness.

Two practical consequences. First, do not drain silken tofu for a blended dessert: the water is part of the intended texture, and removing it yields a denser cream without improving anything. Second, an agar preparation cannot be rescued in the fridge. If it has not set after three hours, the boil was too short or the dose too low, and the whole mixture has to be boiled again with a little more powder.

A silken tofu cheesecake that holds together

For a 20 to 22 cm springform tin, eight slices.

Base: 200 g plain biscuits, 50 g almonds, 60 g melted coconut oil. Grind the biscuits and almonds to a powder, mix with the oil until the texture is sandy and clumps under finger pressure, press firmly into the base of the tin and chill.

Filling: 500 g silken tofu, 200 g plain soy yoghurt alternative, 100 g sugar, 20 g cornflour, the zest of one lemon, a teaspoon of vanilla extract, a pinch of salt. Blend thoroughly, until no grain remains detectable on the tongue.

Baking: 170 °C for 40 to 50 minutes. The centre should still wobble slightly when you move the tin. A cheesecake baked until completely still will be dry once cold, because it keeps setting as it cools.

Resting: at room temperature until lukewarm, then at least four hours in the fridge, ideally overnight. The cornflour and the fat in the base need that time to firm the whole thing up.

The result is less dense than a cream cheese cheesecake, which follows directly from the figures above: there is far less fat to set. Expecting an identical texture leads to adding sugar and fat until the point of the swap disappears entirely.

Soy, isoflavones and what the data show

Soy contains isoflavones, plant compounds whose structure lets them bind to oestrogen receptors, preferentially to the beta receptor. Both the promises and the anxieties start there, usually without any numbers attached.

On menopausal symptoms, a systematic review and meta-analysis published in PeerJ in 2025 retained twelve studies. The overall effect of isoflavones is small: a Hedges’ g of -0.25 (95% confidence interval -0.42 to -0.08) across seven studies and 533 participants. The breakdown is instructive. Significant effects appear for headache, psychosocial symptoms, palpitations and depression, but none for hot flushes, excessive sweating, insomnia or vasomotor symptoms, which is precisely what isoflavones are most often sold for. The authors themselves caution about the small sample sizes.

On breast cancer, the evidence is frankly uncomfortable. Prospective cohort studies associate higher isoflavone intake with reduced risk, but randomised trials looking at markers of risk do not support that protective association. One proposed explanation is that the effect depends on intake during childhood or adolescence, which remains a hypothesis. Association is not causation, and two classes of study that disagree are telling you the answer is not settled.

On infant safety, the Beginnings cohort followed 600 healthy term infants, split evenly between soy formula, cow milk formula and breastfeeding. For skeletal mineralisation, most neurocognitive measures and reproductive organ development, formula-fed infants showed results comparable to breastfed infants.

Applied to a dessert, all of this stays theoretical: one portion delivers a fraction of the intakes studied in that work. Soy in a pudding is a matter of eating well, not a health strategy.

Soy is a declarable allergen

This is the most concrete point on the page if you are cooking for other people. The Swiss ordinance on food information lists soy among the ingredients capable of causing allergies or other adverse reactions. Its articles 10 and 11 require such ingredients to be stated clearly in the ingredient list and to stand out through typeface, font style, background colour or any other suitable means.

That obligation applies to food sold commercially, not to your kitchen. But it indicates the level of care expected: a silken tofu dessert served at a buffet without a word is a problem for someone with a soy allergy, all the more so because a pale vanilla cheesecake looks nothing like soy. Say it out loud, or put a label next to it.

Soy allergy is a separate matter from lactose intolerance, which these desserts genuinely solve. The two get confused often.

Common mistakes

Presenting a silken tofu dessert as high in protein. A slice of cheesecake contains 7.8 g, a flan 7.6 g. Respectable for a dessert, and nowhere near a portion of firm tofu.

Believing a syrup escapes the free sugars category. The NHS files maple and agave exactly where it files white sugar. Choosing a sweetener is a question of taste, not nutrition.

Draining silken tofu before blending it. Its water is part of the texture you are aiming for in a cream or a flan.

Relying on the tofu to set a flan. Agar does the gelling, and only if it has boiled for at least a minute.

Baking the cheesecake until the centre is firm. It finishes setting as it cools. Taken out of the oven motionless, it will be dry.

Serving a soy dessert without mentioning it. Soy is a declarable allergen in food sold commercially, and nothing about a pale cheesecake announces it.

Depending on your situation

You want to lighten a classic dessert. This is silken tofu’s best use, and the effect is dramatic: roughly fifty times less saturated fat than whole cream, volume for volume.

You are after protein. Wrong lever. A portion of firm tofu, legumes or fortified soy yoghurt delivers more. Our page on essential amino acids explains how to spread them across the day.

You are cooking for children. Count free sugars first: 19 g a day from age 4 to 6, which is less than a single slice of the cheesecake above. The fruit creams at 7.7 g suit better, and cutting the syrup improves them further.

You are lactose intolerant. These desserts solve that outright, since they contain no dairy at all. Our page on plant protein myths and realities puts a few received ideas back in their place.

You are hosting and do not know your guests’ allergies. Announce the soy. It is the one precaution these recipes genuinely require.

You want to replace eggs. Silken tofu works in custard-style fillings and much less well where the egg provides structure or aeration. Our guide to egg substitutes handles each case separately.

Storage

Silken tofu is often sold in aseptic packaging, shelf stable at room temperature until opened. Once opened it keeps in the fridge in a closed container and should be used quickly, within two to three days.

Finished desserts keep in the fridge, covered so they do not pick up odours. The cheesecake slices better the next day, once the base has hardened.

Freezing changes silken tofu’s texture, because its water forms crystals that break the protein network. After thawing the texture is spongy. That is acceptable for a preparation you will blend again, and best avoided for a flan that has already set.

The limits of this page

The nutritional values for the recipes are calculated from the composition of the raw ingredients, not measured on the finished desserts. Any given brand of silken tofu, biscuit or yoghurt alternative will drift somewhat from the Swiss database averages, and the portion sizes are calculation conventions rather than prescriptions.

The isoflavone data cover soy intakes well above what an occasional dessert supplies, and often concern supplements rather than foods. They tell you about soy in general, not about a slice of cheesecake.

This page makes no diagnosis. A suspected or confirmed soy allergy is a matter for a clinician, not a recipe adjustment.

Frequently asked questions

Is silken tofu high in protein? No. It supplies 5.0 g per 100 g against 14.7 g for firm tofu. A dessert portion contains between 5 and 8 g.

Can I substitute silken tofu for firm tofu in a recipe? Not in preparations that need to hold their shape when cooked. Silken tofu is 87.6 g water per 100 g and falls apart. For blended fillings the reverse applies: firm tofu gives a grainy texture.

Should I drain it? Not for a blended dessert. The water is part of the texture you want.

Why did my flan not set? The agar probably did not boil long enough. It needs a full minute at a rolling boil to hydrate, and it only gels on cooling below 40 °C.

Is maple syrup better than sugar? In free-sugar terms, no: the NHS puts them in the same category. The syrup brings a different flavour and a little calcium, which does not change the nutritional comparison.

Are these desserts suitable for children? The fruit creams yes, with the sweetener reduced. A single slice of the cheesecake above exceeds the daily free-sugar limit for a child aged 4 to 6.

Does soy disrupt hormones? Isoflavones bind to oestrogen receptors, but the measured effect on menopausal symptoms is small and absent for hot flushes. At dessert quantities the question is largely theoretical.

Do I need to flag the soy when serving these? Yes. Soy is a declarable allergen in food sold commercially, and nothing about a cheesecake’s appearance announces it.

Sources

• Swiss Food Composition Database, FSVO
• FSVO, Swiss dietary recommendations
• NHS, Sugar: the facts
• Swiss ordinance on food information (SR 817.022.16)
• PeerJ 2025, soy isoflavones and menopausal symptoms
• Nutrients 2026, early soy intake and breast cancer
• Advances in Nutrition 2026, Beginnings cohort, soy-based formula
• Food Chemistry X 2023, tofu coagulation with calcium sulfate and glucono-delta-lactone

To go further, see how to read a nutrition label and the guide to marinating tofu and tempeh.