Most people who start counting protein discover the same thing in the first week: the number is lower than they expected. Sometimes dramatically. A client of mine who was certain she was eating "loads of protein" as a vegetarian came in at 48 grams on her first tracked day, against a target of 110.
The interesting part is what happened next. Within three weeks she was consistently clearing 115 grams without adding a single supplement, without eating anything she disliked, and without her grocery bill changing much. Nothing about her food preferences changed. What changed was that she could see the number.
That is what a protein counter app is for. But counting is a skill with its own failure modes, and a counted number can be wrong in several predictable ways. This guide covers where counting goes astray, how to correct it, and why counting protein on a plant-based diet requires a slightly different approach than the standard advice assumes.
Counter or Tracker: Is There a Difference?
Practically, no. "Protein counter app" and "protein tracker app" describe the same category, and app developers use the terms interchangeably. If there is a shade of difference, "counting" implies the arithmetic of adding up grams, while "tracking" implies watching the number over time. A good app does both.
What genuinely distinguishes apps is scope. Some count protein only. Others count protein alongside calories, carbohydrates, and fat. The narrower ones are considerably faster to use and, in my experience with clients, far more likely to still be in use after a month. If your goal is protein, count protein.
Where Protein Counting Goes Wrong
Six sources of error account for nearly all the inaccuracy I see. Most are easy to correct once you know they exist.
Cooked versus raw weight
The single most common mistake. Meat loses water during cooking and therefore concentrates: roughly 150 grams of raw chicken breast becomes about 110 grams cooked, containing the same protein. If you weigh it raw and log the cooked database entry, you overcount by around thirty percent.
Rice and pasta run in the opposite direction, absorbing water and becoming less protein-dense by weight. Dried pasta at around 12 grams of protein per 100 grams becomes roughly 5 grams per 100 grams once cooked.
The fix is simple: check whether the database entry says raw or cooked, and match your weighing to it. Pick one convention and hold to it.
Label rounding
Nutrition labels are permitted to round, and small roundings accumulate. A protein bar listing 20 grams might contain 19.5. A yogurt listing 10 might contain 9.6. Across a day of packaged foods this can amount to several grams. Not worth agonising over, but it is one reason your true intake tends to sit slightly below your logged intake rather than above it.
Serving size assumptions
Database entries default to a serving that may bear no relation to what you ate. "1 container" of yogurt could be 150 grams or 227 grams. "1 slice" of bread varies by more than double between brands. Whenever you log something new, look at what the app thinks a serving is before accepting it.
Crowd-sourced database errors
Apps that let users add entries accumulate duplicates and mistakes at an impressive rate. You will find the same product listed four times with four different values, at least one of which has a decimal point in the wrong place. Prefer generic entries for whole foods, and verify branded entries against the physical label the first time you log them.
Uncounted protein in non-protein foods
This one runs in your favour and is the most common surprise, particularly for vegetarians. Bread, oats, rice, pasta, vegetables, nuts, and seeds all contain protein. Individually the amounts look trivial. Collectively they are frequently 25 to 40 grams a day.
A realistic example. Someone eats oats with milk for breakfast, a sandwich at lunch, pasta with vegetables at dinner, and hummus with a wholemeal pitta as a snack. They log "no real protein today". In reality:
- 80 g dry oats: ~11 g
- 250 ml milk: ~8 g
- 2 slices wholemeal bread: ~8 g
- 100 g dry pasta: ~12 g
- Mixed vegetables: ~5 g
- 60 g hummus with a pitta: ~9 g
That is 53 grams before a single deliberate protein source. Apps that only prompt you to log obvious protein foods will systematically undercount days like this. Log everything you eat, or at least everything substantial.
Cooking losses and additions
Oil added during cooking changes calories, not protein. But protein does leach into cooking water when boiling, and fat trimmed from meat removes a little protein along with it. These effects are small — a few percent — and not worth modelling.
Counting Protein on a Plant-Based Diet
The standard framing, that plant-based eaters struggle to get enough protein, is largely wrong. Most vegetarians and vegans in developed countries meet the RDA comfortably. What is true is that reaching the higher targets associated with resistance training or muscle preservation — 1.6 g/kg and above — genuinely takes deliberate planning on a plant-based diet, and counting works differently.
Protein is distributed rather than concentrated. An omnivore hits 120 grams with three obvious sources: eggs, chicken, fish. A plant-based eater hits the same total with twelve smaller contributions across the day. This makes mental estimation much harder and makes an app correspondingly more valuable.
The adjustment for quality. Most plant proteins are somewhat less digestible and have less favourable amino acid profiles than animal proteins. The standard practical correction is to aim roughly 10 to 20 percent higher than the equivalent omnivorous target. Someone who would target 120 grams on a mixed diet should aim for 135 to 145 grams fully plant-based.
Combining within a meal is unnecessary. The old advice about pairing rice with beans at every meal has not survived scrutiny. The body maintains a free amino acid pool that buffers across meals, so variety across the day is sufficient. Eat a range of legumes, grains, soy foods, nuts, and seeds and the amino acid profile takes care of itself.
Soy deserves particular mention. Soy is the one widely available plant protein that scores comparably to animal sources on amino acid completeness. Tofu, tempeh, edamame, and soy milk make hitting higher targets substantially easier, and the concerns about soy that circulate online are not supported by the evidence in humans at normal dietary intakes.
Plant protein values worth knowing
| Food | Portion | Protein |
|---|---|---|
| Seitan | 100 g | ~25 g |
| Tempeh | 100 g | ~19 g |
| Lentils, cooked | 1 cup | ~18 g |
| Firm tofu | 100 g | ~17 g |
| Edamame, shelled | 1 cup | ~17 g |
| Chickpeas, cooked | 1 cup | ~15 g |
| Black beans, cooked | 1 cup | ~15 g |
| Soy milk | 1 cup | ~7 g |
| Hemp seeds | 30 g | ~10 g |
| Peanut butter | 2 tbsp | ~8 g |
| Quinoa, cooked | 1 cup | ~8 g |
| Oats, dry | 40 g | ~5 g |
| Wholemeal bread | 1 slice | ~4 g |
The reliable structure for a plant-based day is a substantial legume or soy source at two meals, a grain at each meal, and nuts or seeds as snacks. That framework clears 100 grams without difficulty.
Making Counting Fast Enough to Sustain
Accuracy is worthless if the process is so tedious you abandon it. A few things make the difference.
Pin your regulars in week one. Almost every counter app supports favourites. Fifteen pinned foods usually covers eighty percent of what a person eats. This turns a thirty-second search into a single tap.
Build a handful of composite entries. If you eat the same overnight oats four mornings a week, create it once as a custom food rather than logging five components each time. This is often a paid feature, and it is one of the few genuinely worth paying for if your diet is repetitive.
Weigh for one week, then estimate. Portion estimation is the biggest error source and the easiest to fix. A week with a kitchen scale recalibrates your eye permanently. After that, estimate.
Use photo scanning only for genuinely unknown meals. Features like AI Meal Scan are valuable when you are handed a plate you did not prepare. For your own repeated meals, tapping a favourite is faster and more accurate. Photo estimation cannot see cooking oil, cannot judge sauce composition, and cannot weigh a portion precisely — review the estimate before accepting it.
Do not chase the last five grams. Landing within about ten percent of your target most days produces the same result as hitting it exactly. Precision beyond that is a hobby, not a requirement.
What to Do With the Number
Counting produces data. Data on its own changes nothing. Three things to actually look at:
Your weekly average. Single days fluctuate for reasons that have nothing to do with your habits. The weekly average is what relates to outcomes.
Your breakfast. In nearly every client I have worked with, breakfast is the weakest meal by a wide margin — frequently under 10 grams. It is also the easiest to fix. Moving from toast to Greek yogurt, tofu scramble, or overnight oats with soy milk and hemp seeds adds twenty grams to a meal that had almost none, and it does more for your daily total than any other single change.
Your weekend. Most people's tracked weekdays look nothing like their untracked weekends. If you only log Monday to Friday, you are measuring your best days and drawing conclusions about your average.
Frequently Asked Questions
What is the difference between a protein counter and a protein tracker?
Nothing meaningful. The terms are used interchangeably for apps that add up your daily protein intake. What matters is whether the app counts protein alone or counts full macros — single-nutrient apps are faster to use and have much better long-term retention.
Do I need to count the protein in bread, rice, and vegetables?
Yes, if you want an accurate number. Non-obvious sources typically contribute 25 to 40 grams a day, which is a substantial fraction of most targets. Apps that only prompt you to log meat, dairy, and supplements will systematically undercount, and this affects vegetarians most of all.
Should I weigh food raw or cooked?
Either, as long as it matches the database entry. Meat loses water and concentrates when cooked — 150 g raw chicken is roughly 110 g cooked with the same protein — while grains and pasta absorb water and dilute. Mismatching the two is the most common counting error there is.
How much protein do I need on a vegan diet?
Take your standard target and add roughly 10 to 20 percent to account for lower digestibility and amino acid profiles. For someone training regularly, that typically means 1.8 to 2.0 g/kg rather than 1.6. Soy foods, seitan, and legumes make this achievable without supplements, though a protein powder can simplify busy days.
Is it necessary to combine plant proteins at each meal?
No. The idea that rice and beans must be eaten together at the same sitting has been superseded. Your body maintains a circulating pool of amino acids that buffers across meals, so eating varied plant protein sources over the course of a day is sufficient.
How accurate are protein counter apps really?
The arithmetic is exact; the inputs are estimates. Generic entries for whole foods are reliable, branded crowd-sourced entries frequently are not, and portion estimation introduces the largest error of all. Expect to be within roughly ten percent of your true intake with reasonable care, which is more than accurate enough for any practical goal.
Try Protein Tracker
Protein Tracker keeps counting fast: a curated database of over 100 high-protein foods including plant sources, one-tap favourites for the meals you repeat, a daily ring showing grams remaining, and weekly averages so you can see the trend rather than the noise.
The free tier includes one protein log per day. Pro adds unlimited logging, custom foods for your own recipes, AI Meal Scan for meals you did not cook, and editing of past entries, available weekly or as a one-time lifetime purchase.
Your logs stay on your device by default.
