FitBeyond AI

Sources

Last updated: 15 October 2026

FitBeyond AI turns a few facts about your body and your goal into a daily calorie target, macro targets, and a plan with a date on it. This page lists every recommendation the app makes, the exact number it uses, and the published source behind it.

Some of these numbers are settled science. Others are practical rules of thumb that the research qualifies or contradicts. We have listed both kinds here, and where our number and the evidence disagree, we say so rather than leaving it out.

FitBeyond AI is a general wellness app, not a medical service. The targets it calculates are estimates meant to guide you, not medical, dietary, or clinical advice, and the app is not intended to diagnose, treat, cure, or prevent any condition. Talk to a qualified healthcare professional before making significant changes to your diet or exercise, and especially if you are pregnant, breastfeeding, or managing a medical condition.

Your daily calorie target

Energy at rest

The app estimates the energy your body uses at rest with the Mifflin-St Jeor equation, from your weight, height, age, and gender. It calculates 10 × weight in kg, plus 6.25 × height in cm, minus 5 × age, then adds 5 for men and subtracts 161 for women. If you would rather not specify, it uses 78, the midpoint of the two.

Frankenfield and colleagues compared the four equations in common clinical use and found Mifflin-St Jeor landed within 10 percent of measured resting metabolism more often than any of the others, with the narrowest error range. They also found meaningful error at the level of an individual person, which is why the app treats the result as a starting point you can edit rather than a measurement of you.

Your activity level

That resting figure is multiplied by an activity factor drawn from how often you train each week. The app has five factors: 1.2, 1.375, 1.55, 1.725, and 1.9. The workouts question in onboarding offers three bands, and they map to 1.375 for nought to two workouts a week, 1.55 for three to five, and 1.725 for six or more. The other two only apply to a plan that carries a workout count without a band: 1.2 when no workouts at all are recorded, and 1.9 for eight or more a week.

Multiplying resting energy by a physical activity level is the method set out by the FAO, WHO, and UNU expert consultation on human energy requirements. Our factors follow the convention used across clinical calculators rather than that report's own bands. Worth knowing: the report's sedentary category starts at 1.40, so the 1.375 we apply to the lowest band is a shade below its lowest published figure, and the 1.2 used when no workouts are recorded sits further below it. Our 1.725 and 1.9 both sit inside its active range, which runs from 1.70 to 1.99.

When you tell us about your day and your workouts

If you answer the Activity questions on the You tab and accept the plan change it proposes, the app swaps the single workouts-per-week band above for two separate pieces: how active your day is outside workouts, and the workouts you do on top.

Your day outside workouts sets a physical activity level. We use the Dietary Reference Intakes' categories as the anchors, roughly 1.4 for inactive, 1.6 for low active, and 1.75 for active, and the FAO / WHO / UNU report's table of activity levels for work that keeps you on your feet or is physically demanding, where moderately active construction work sits between 1.70 and 1.99. That report also notes that a level above 2.40 is difficult to maintain over time, so the app never lets the combined level go past it, however much you train.

Workouts are then added on top of that level. Each session adds the energy of its minutes above rest, using the Institute of Medicine's method: 0.0175 kcal per kg per minute for each MET above rest, raised by 1.15 for the energy the body spends recovering and divided by 0.9 for the energy of digesting food, which comes to 1.34 kcal per kg of body weight per hour for each MET above rest. The report writes the same thing as a change in activity level, (METs minus 1) × 1.34 × minutes ÷ 1,440 for its reference man. Effort maps to MET values of 3.5 for easy, 5.0 for moderate, and 7.5 for hard, taken from the 2024 Compendium. Session lengths are counted at the middle of the band you pick, 15, 30, 50, or 75 minutes.

Daily step ranges, if you give one, are used only as a cross-check on the level above, moving it by one step at most, and never replace it. The pedometer bands in Tudor-Locke and Bassett's review are a population-level guide to how active a day is, not a measurement of your energy use. Like every figure on this page, the result is an estimate you can edit.

The answers can change protein too. If strength training is one of your workouts, protein moves to 2.2 g/kg, following the sports nutrition position stand and the Morton meta-analysis cited below. Morton and colleagues found the benefit for lean mass levelled off around 1.6 g/kg, with a confidence interval reaching 2.2, so the app uses the top of that interval for people who lift. Protein still never takes more than 35 percent of your calories.

Turning a goal into a daily deficit

To convert a weekly pace into a daily calorie change, the app uses roughly 7,700 kcal per kilogram of body mass, divided across seven days. That figure traces back to Wishnofsky's 1958 estimate of the energy stored in a pound of body fat, about 3,500 kcal, and it is the arithmetic behind nearly every calorie tracker.

It is a planning approximation, not a promise. Your metabolism adapts as you lose weight, so real loss runs slower than this straight-line rule predicts, and the gap widens the longer you diet. Hall and colleagues showed body weight responds to a change in intake with a half-time closer to a year, and Thomas and colleagues found the rule over-predicted 24-week weight loss by around 11 kg where a validated dynamic model was off by 2.2 kg. We use the simple rule because it makes a plan concrete on day one. We would rather tell you its limits than imply a precision it does not have. If you want a model that accounts for adaptation, the NIH Body Weight Planner is free and public.

How fast the app lets you go, and how low it lets you eat

The app caps your weekly pace at 1 percent of your body weight, and never more than 1 kg a week whatever you weigh. The Recommended badge points at half of that, 0.5 percent a week. Ask for a faster pace and the app labels it as faster than recommended rather than hiding it. Whatever pace you choose, it never prescribes fewer than 1,200 calories a day, and the same floor applies when you edit a target by hand.

Clinical guidance puts a sustainable rate at 1 to 2 pounds, roughly 0.45 to 0.9 kg, per week, produced by a deficit of 500 to 1,000 calories a day, with an initial goal of losing about 10 percent of starting weight. The same guidance places structured low-calorie diets at 1,000 to 1,200 calories for women and 1,200 to 1,500 for men, under professional supervision. Our floor sits at the top of that supervised range, so the app on its own will never take you below it.

Body mass index

The BMI scale in the app uses the World Health Organization's adult bands: below 18.5 is underweight, 18.5 to 24.9 is healthy, 25 to 29.9 is overweight, and 30 or above is obese. That same 18.5 threshold is the floor the app clamps a goal weight to, so you cannot set a target that would place you in the underweight range for your height.

BMI is a rough guide. WHO treats it as a surrogate marker of body fatness rather than a diagnosis, and it cannot tell muscle from fat, so it can mislead for athletes, during pregnancy, and in other cases. A health professional can give you the full picture.

Protein, fat, and carbohydrate

Protein scales with your goal and with how often you train, rather than with body weight alone. Losing weight and rarely training is set at 1.4 grams per kilogram; losing weight and training six or more times a week is set at 2.0. Maintaining lowers the figure and gaining sits between the two. A deficit raises it at every activity level because eating less, not training, is what makes protein most worth protecting: it is a large part of what decides whether the weight you lose comes off as fat or as muscle.

The recommended dietary allowance is 0.8 g/kg, the amount that avoids deficiency rather than the amount that supports a goal, and no target we calculate goes below it. Sports nutrition guidance supports 1.4 to 2.0 g/kg for people who exercise, and a meta-analysis of 49 trials found gains in lean mass plateau near 1.6 g/kg with resistance training, which is why only the most active bands reach the top of that range.

Two guards sit on top. The coefficient is applied to a weight inside the healthy BMI range for your height rather than to your current weight, so the target scales with the body rather than with body fat. And protein never claims more than 35 percent of your calories, so a low target cannot end up as a plan that is mostly meat. From age 65 the figure is held at a minimum of 1.2 g/kg, since holding on to muscle gets harder with age. That follows the PROT-AGE expert group, which recommends at least 1.0 to 1.2 g/kg a day for people over 65, and 1.2 or more for those who exercise, with an exception for people with severe kidney disease. The app has its own kidney setting: telling it about a kidney condition sets protein at 0.8 g/kg, following the kidney guideline listed under Health and body below.

An earlier version of the app set protein at a flat 2 g/kg of current body weight for everybody. That number came from the top of a range this same page scopes to people who exercise, and applying it to someone who does not train, on their current weight rather than a healthy-range one, produced targets well above anything the research here supports. It has been replaced by the bands above.

Fat is set at 25 percent of your calories, inside the 20 to 35 percent range the Institute of Medicine gives for adults, and carbohydrate takes whatever calories remain. A day's fat only counts as inside its budget from 20 percent of your calories up, the bottom of that range. Grams become calories with the Atwater factors: 4 kcal per gram of protein and of carbohydrate, 9 kcal per gram of fat. These general factors were derived for mixed diets and are less accurate applied to a single high-fiber food.

Because carbohydrate is a remainder, it is not held to the carbohydrate references, and here our number and the evidence can part ways. The Institute of Medicine's range for carbohydrate is 45 to 65 percent of calories, and its recommended dietary allowance is 130 grams a day. On a low-calorie plan where protein reaches its 35 percent cap, carbohydrate lands at about 40 percent of calories, and at the 1,200 calorie floor at about 120 grams, under both. The app says so where it explains your carb number.

If you tell the app in Food preferences that you eat low-carb or keto, it offers a lower-carbohydrate split as a plan change you can accept or decline. Carbohydrate is held to the larger of 35 percent of your calories and 130 grams, the recommended dietary allowance, and fat takes the calories that frees up. The American Diabetes Association's consensus report counts 26 to 45 percent of energy from carbohydrate as low carbohydrate, so the split sits at the gentle end of that range. The app never sets a ketogenic target. If you take medicine for diabetes, talk to your doctor before cutting carbohydrate, since a dose may need to change.

Fiber, sugar, and sodium

Fiber is set at 14 grams per 1,000 calories of your target, the Institute of Medicine's adequate intake, and never shown as less than 20 grams a day. That 20 gram floor is our own rather than a published figure. The Institute's rate gives 17 grams at the app's 1,200 calorie floor, while the European Food Safety Authority and WHO both put adequate fiber for adults at 25 grams a day. Reaching any of these is worth something: the Dietary Guidelines for Americans report that more than 90 percent of women and 97 percent of men fall short of the recommended intake.

The sugar limit is 10 percent of your calorie target, counted as added sugar. WHO's strong recommendation is to keep free sugars below 10 percent of energy, with a conditional recommendation that going below 5 percent brings further benefit, and the FDA's Daily Value for added sugars is the same 10 percent, 50 grams on a 2,000 calorie diet. One distinction worth knowing: WHO's "free sugars" is broader than "added sugars", because it also counts the sugars naturally present in honey, syrups, and fruit juice.

Here our number and our data do not line up exactly, and it is worth saying so. The app counts the added sugar a meal's analysis estimates or a food's label reports, so the sugar in whole fruit and the lactose in milk, which the guideline is not aimed at, stay out of it. Two gaps remain. The sugar in fruit juice counts toward WHO's limit but is not added sugar, so a day with a lot of juice can read lower than WHO would count it. And a meal that carries no added-sugar figure at all, such as one logged before the app recorded it or a food whose database entry leaves it out, counts its total sugars instead, fruit and milk included, so that day can read higher. Treat this as the softest number on the page.

Sodium is a flat 2,300 mg a day. That is the National Academies' Chronic Disease Risk Reduction intake for adults, the level above which eating less is expected to lower the risk of high blood pressure and heart disease, and it is also the Daily Value on US nutrition labels. It is a limit to stay under over time, not a toxicity threshold, and a day under it is never a miss: the adequate intake is 1,500 mg. WHO and the European Food Safety Authority set a lower figure for adults, 2,000 mg a day. For context, average intake in the United States runs closer to 3,400 mg, and more than 70 percent of it comes from packaged and prepared food rather than the salt shaker.

Water

The app sets a daily water target of about 35 ml per kilogram of body weight, held between 1.5 and 4 litres, counted against the drinks you log. This is a practical scaling rule of our own, and we want to be straightforward that it is a heuristic rather than a clinical requirement.

The reference figures are flat daily volumes, not per-kilogram ones, and they differ by sex: the Institute of Medicine's adequate intakes are 3.7 litres a day for men and 2.7 litres for women, counting all drinks plus the water in food, of which drinks supply roughly 80 percent. The European Food Safety Authority's are 2.5 and 2.0 litres. There is no established per-kilogram figure for adults, and our target does not yet differ by sex: for a 70 kg person it asks for 2.45 litres of drinks, between the drink share of the two Institute of Medicine figures.

The Institute is explicit that its adequate intake is not a requirement, and that for a healthy person a day below it may carry no extra risk. That is why the app never treats a day under its water target as a miss. Real needs move with activity, heat, and what you eat, and for healthy people thirst remains the primary signal.

Calories burned in a workout

When you log exercise, the app estimates the burn as the activity's MET value, scaled by how hard it felt, multiplied by your body weight and the duration in hours. One MET is defined as an energy cost of 3.5 ml of oxygen per kg per minute, which is about 1 kcal per kg per hour, and that definition is exactly the arithmetic the app uses. The per-activity values it starts from are the Compendium's representative figures for moderate effort, for example 3.5 for walking, 7.5 for cycling, and 9.8 for running.

The Compendium's authors are explicit that MET values do not produce a precise number for an individual, because resting metabolism varies with age, sex, size, and body composition. Treat a logged burn as an estimate, and edit it if you know better. The app always lets you type your own figure over ours.

Why the app asks you to log

Reminders, streaks, and the nudge to log a meal are not decoration. Keeping a record of what you eat is one of the better-supported behaviours in weight management. A systematic review of the literature found dietary self-monitoring consistently associated with weight loss, and a later study of electronic logging found that how often people logged, rather than how long they spent doing it, was what tracked with success. Both are observational findings, so read them as a strong association rather than a guarantee, and the CDC recommends a food diary on the same basis.

Health and body

When you tell Calby about a pregnancy, a health condition or a medicine, these are the guidelines your plan follows. Your doctor's advice always comes first.

Corrections

If you think we have cited something wrongly, drawn too much from a source, or missed research that contradicts a number above, please tell us at support@fitbeyond.app. We would rather fix it than defend it.