Tofu Stir-Fry

What this is

Tofu, vegetables and a savoury sauce, cooked hot and fast. Twenty minutes.

I put this on the list because tofu is the food people write off fastest, usually on the basis that plant protein is inferior. That claim rests on a specific technical measure, and the relationship between that measure and what actually happens to people has been examined properly.

The answer is more interesting than either side of the usual argument.

Ingredients

For 2 servings

400 g firm or extra-firm tofu, pressed 2 tbsp cornflour 3 tbsp neutral oil 1 head broccoli, small florets 1 red pepper, sliced 1 carrot, sliced thin on the diagonal 3 spring onions, sliced 3 cloves garlic, minced 1 thumb ginger, minced

Sauce 3 tbsp soy sauce 1 tbsp rice vinegar 1 tbsp maple syrup or honey 1 tsp sesame oil 1 tsp cornflour 100 ml water Chilli flakes to taste

To serve: rice, sesame seeds, coriander

Method

1. Press the tofu properly. Wrap it in a clean towel, put a heavy pan on top, and leave it 20 minutes. This is the step that decides whether your tofu is good or wet, and it cannot be skipped.

2. Cut into 2 cm cubes and toss with the cornflour until lightly coated all over.

3. Whisk the sauce ingredients together, including the cornflour, until smooth.

4. Heat 2 tbsp oil in a wok or wide pan over high heat. Fry the tofu in a single layer without moving it for 2 minutes, then turn and repeat until golden on most sides, about 6 to 8 minutes total. Remove and set aside.

5. Add the last of the oil, then the harder vegetables, carrot and broccoli, and stir-fry 3 minutes. Add the pepper for another 2.

6. Add the garlic, ginger and spring onion whites, 30 seconds.

7. Return the tofu, pour in the sauce, and toss for 1 to 2 minutes until it thickens and coats everything.

8. Finish with spring onion greens, sesame seeds and coriander over rice.

The protein quality paradox

This is the section worth reading even if you never cook tofu.

A review examined the role and limitations of protein quality metrics, particularly the Digestible Indispensable Amino Acid Score, when applied beyond simply assessing amino acid adequacy.

Its finding is stated directly, and the title puts it as a paradox. Although animal-derived protein sources typically achieve higher amino acid-based quality scores, recent evidence from randomised controlled trials and large-scale prospective cohort studies indicates these differences do not consistently translate into superior cardiometabolic or mortality-related outcomes.

A higher score that does not consistently produce a better outcome.

It goes further. Substitution analyses and intervention studies suggest that replacing animal-based proteins with plant-based sources is associated with favourable changes in cardiovascular risk markers and reduced risk of chronic disease.

The authors attribute this to the food matrix, meaning the bioactive compounds and everything else that arrives alongside the protein, and they situate all of it within overall dietary patterns rather than isolated nutrients.

The crucial qualifier, which they state and I am repeating: this concerns populations with overall adequate protein intake. If you are not hitting your protein needs at all, source quality matters a great deal more.

So the honest position is not that plant protein is secretly superior. It is that in people already eating enough protein, the quality score that gets used to rank sources is a poor predictor of the outcomes people actually care about.

This is a narrative review synthesising trials and cohorts, not a trial itself, and it addresses long-term health outcomes rather than muscle growth in trained lifters.

What the long-term trials show

For the muscle question specifically, there is a meta-analysis that set a high bar.

It assessed the long-term effects of plant-based protein compared with animal-based protein supplementation on body composition, muscle strength, physical performance and cardiometabolic risk factors in adults aged 18 and over. PubMed, Scopus and Web of Science were searched from inception to 25 February 2025, with citation tracing, and only randomised controlled trials comparing the two for at least six months were included. Data were pooled with a random effects model.

Six months minimum. That inclusion criterion is why this is worth citing: most protein source comparisons run eight to twelve weeks, which is long enough to see nothing and conclude anything.

The review exists because previous studies yielded mixed results, and the authors wanted pooled long-term evidence.

I am reporting this study's design and rationale rather than its pooled effect estimates, which are the substance of the paper and not something to compress from an abstract.

What you can take from it: the question of whether plant protein supplementation matches animal protein over six months or more has been assembled properly, which is more than can be said for most of the arguments you will hear about it.

This is a meta-analysis of supplementation trials using protein isolates and concentrates rather than whole foods like tofu, and its effect estimates are not reported here.

Why press the tofu, and why the coating

There is a mechanical reason behind both instructions, and a research angle on the second.

A study evaluated the digestibility and Digestible Indispensable Amino Acid Score of 17 plant-based substrates using an internationally harmonised in vitro protocol, comparing results against standardised ileal digestibility measured in pigs on the same substrates. Except for corn flakes, the in vitro protocol predicted crude protein and amino acid digestibility with absolute errors below 20 percent.

The part that matters for cooking: the effect of heat treatment on rapeseed protein digestibility was evidenced by the peptide profile, pointing to a specific protein as the main contributor to protein resistance. The study also notes that trypsin inhibitors in a sample can limit these methods.

Heat treatment changing how digestible a plant protein is, and specific proteins resisting digestion, is the general principle behind why plant foods are usually cooked rather than eaten raw.

That study examined rapeseed and 16 other substrates, not tofu, and its purpose was validating a laboratory method against pig data. I am carrying the principle about heat and plant protein digestibility, not a claim about your wok.

The pressing and the cornflour are pure technique: water leaving the tofu is what stops it browning, and the starch coating gives the surface something to crisp. Neither is a nutrition claim.

Storage and swaps

Keeps 3 days in the fridge. The tofu softens as it sits, so it is best on day one, and reheating in a hot pan restores more texture than a microwave.

Swap tofu for tempeh, which is firmer and fermented, or for edamame if you want it faster.

Swap the vegetables for whatever is in the fridge. This recipe is a template, not a prescription; harder vegetables go in first, softer ones later.

Add cashews or peanuts at the end for fat and crunch.

And get the pan properly hot before the tofu goes in. A lukewarm pan gives you steamed tofu, which is the outcome everybody who dislikes tofu has actually experienced.