Nutrition

Say you are a 32-year-old man, 178 cm, 82 kg, and you have not had a body analysis yet.

The engine computes your basal metabolism with the Mifflin-St Jeor equation: 10 × weight + 6.25 × height − 5 × age + 5.

There are two formulas and the choice turns on one question: is your lean body mass known. If it is, the Cunningham equation runs (500 + 22 × lean mass), because muscle burns more energy than fat and knowing lean mass narrows the estimate. If not, Mifflin-St Jeor is used from weight, height and age. For women the final term is −161 rather than +5.

Once you do a body analysis and give the system your lean mass, your calorie target rests on a more personal base.

The system calls this the "BMR estimate".

Say your job is moderately active, you walk 6–10 thousand steps a day and you train four days a week.

Your basal metabolism is multiplied by a factor: moderately active work gives a base of 1.35, the 6–10 thousand step band adds 0.05, and each weekly session adds 0.025 — 1.50 in total.

Basal metabolism is what you would burn without moving at all; living the day sits on top of it. The factor is clamped between 1.0 and 2.0, because left unbounded someone declaring "20 thousand steps a day plus 7 training days" would arrive at an unrealistic target.

Declare your step band and weekly training days honestly — these two fields directly enlarge or shrink your target.

The system calls this "total daily energy expenditure" (TDEE).

Say your daily expenditure was computed at 2,666 kcal and your goal is fat loss.

15% is subtracted from your expenditure — 2,666 × 0.85 = 2,266 kcal — and the result is checked against a floor.

The delta is a percentage of expenditure, not a fixed number of calories: cutting 500 kcal is measured for someone burning 3,000 and dangerous for someone burning 1,600. Muscle gain applies +10%, strength and endurance +5%, recomposition −5%. The result cannot go below 1,500 kcal for men and 1,200 for women; that floor rests on the ACSM 2021 position stand, below which meeting micronutrient needs becomes difficult.

When you change your goal your calorie target is recomputed the same day; if the floor kicked in, losing faster means moving more rather than cutting calories.

The system calls this the "goal delta plus calorie floor".

Say your daily target is 2,266 kcal and you weigh 82 kg; the screen reads 180 g protein, 66 g fat, 238 g carbohydrate.

Protein is set first per kilogram (2.2 g/kg on fat loss → 180 g), then fat (0.8 g/kg → 66 g), and the remaining calories go to carbohydrate: (2266 − 180×4 − 66×9) ÷ 4 = 238 g.

The order is not arbitrary. In a deficit protein limits muscle loss, so it is the first macro protected; fat has hormonal functions, so it is never taken below 30 grams a day; carbohydrate is the most flexible, so it carries the remainder. The protein ceiling is 300 g and the carbohydrate floor 50 g. These thresholds come from the ISSN 2017 position stand and are written in the code with their citation.

As your weight changes your protein and fat grams change and carbohydrate is recomputed from the remainder — you do not need to track all three separately.

The system calls this the "macro policy".

Say your body mass index is above 30 and multiplying by your raw weight would have pushed your protein target past 250 grams.

In that case the engine uses a calculation weight: ideal weight plus a quarter of the excess, instead of your raw weight.

Fat tissue creates no protein requirement; multiplying by raw weight overstates the need, and the unnecessary protein eats the calories left for carbohydrate. The source is again ISSN 2017 together with Choban & Dickerson.

At a high body weight your protein target will not look proportional to your raw weight — that is a measured correction, not an error.

The system calls this the "adjusted body weight".

Say Tuesday is a training day and Wednesday a rest day. Your daily calorie target is 2,266 kcal on both.

The calorie total stays the same on both days; what changes is the gram split between carbohydrate and protein. On a training day carbohydrate rises with intensity — +20% at high, +15% at moderate, +10% at low — and falls 10% on a rest day.

Carbohydrate is training fuel; raising it on the day you need it and lowering it on the day you do not preserves performance within the same weekly calories. For those training five or more days a week, protein rises a further 0.1 gram per kilogram. A training day also enlarges the snack’s share by 0.10 while a rest day shrinks it by 0.02 — because an extra fuelling window opens around training.

When you change your training plan, your nutrition plan redistributes across the week too; you do not have to tune the two modules separately.

The system calls this "carb cycling".

Say your daily target is 2,266 kcal and you eat four meals; the breakfast card reads "30% · ~680 kcal".

The daily calories are first split into meal shares (breakfast 0.30, lunch 0.35, dinner 0.30, the remainder to the snack), then foods that hit each meal’s calorie target are selected from the catalogue.

The selection is scored rather than random: candidate foods are ranked by how close they come to that meal’s calorie and macro target, repetition of the same food within the week is limited, and the winner is decided deterministically. Determinism means the same profile produces the same plan — if the plan changed on every open, neither a shopping list nor weekly tracking would be possible.

The foods in your plan come out of your target and the catalogue’s macros; swapping an item you dislike for its alternative does not disturb the day.

The system calls this the "week plan builder".

Say you declared a gluten intolerance and see no wheat products in your plan.

Your declaration is converted into an allergen tag and every food carrying that tag is removed from the pool before the plan is built.

The filtering runs BEFORE the plan builder, not after. Filtered afterwards, the builder would compose a meal with that food, the item would then be removed and the meal would fall below its calorie target. Foods you dislike work differently: individually chosen foods are removed entirely, while disliked tags are not removed but merely down-scored — because deleting a whole tag could starve the pool.

If you definitely do not want a food, select it individually; saying "I do not like this style" pushes it back rather than removing it.

The system calls this the "food constraints filter".

Say you have followed your plan for three weeks but the scale has not moved.

The engine re-measures your target: with good adherence it cuts your calorie target by 3%; with poor adherence it leaves the target alone.

The engine separates a stall into two causes — metabolic adaptation or not following the plan — and their remedies are opposite. Cutting calories without measuring adherence would needlessly lower the target of someone who simply is not following the plan, widening the deficit. If loss exceeds 1% of body weight a week, the target is raised 5% instead. Total drift from the starting target is capped at 15%.

When the scale stalls, first get your meal ticks in order — the engine’s ability to decide correctly depends on them.

The system calls this "dynamic TDEE adaptation".