On This Page
Important — Not a Medical Device. MacroLeaf AI is a general wellness and nutrition tracking app. It is not a medical device, has no FDA / CE / regulatory clearance, and does not provide medical diagnosis, treatment, or advice. All calculations and recommendations are estimates based on the published peer-reviewed sources below. Always consult a licensed physician, registered dietitian, or other qualified healthcare professional before making medical decisions, starting a diet, or changing your treatment.
How We Use These Sources
- All food nutrition values come from validated public databases (USDA, OFF, ICMR).
- All formulas (BMR, TDEE, macros, hydration) come from peer-reviewed equations.
- Weight-change estimates use published energy-balance models.
- AI Coach replies are bounded by hardcoded safety rules and medical refusal logic.
- Photo and voice estimates are explicitly labeled as approximations.
1.Food Nutrition Data
Per-food calorie and macronutrient values shown in the Diary, Meals tab, search results, and AI Coach come from these public, peer-validated databases:
USDA FoodData Central
U.S. Department of Agriculture, Agricultural Research Service. FoodData Central. fdc.nal.usda.gov. 2019-present.
Primary source for raw foods, packaged products, and Survey (FNDDS) entries used in the United States. Each food's calories, protein, carbohydrate, fat, fiber, and sugar values are pulled directly from FDC's API.
https://fdc.nal.usda.gov
Open Food Facts
Open Food Facts contributors. Open database of packaged food products. world.openfoodfacts.org. 2012-present.
Used for barcode lookups and packaged products that are not in USDA. Crowd-sourced but verified against label data; nutrition values are normalized to per-100g.
https://world.openfoodfacts.org
Indian Food Composition Tables (IFCT)
Longvah T., Ananthan R., Bhaskarachary K., Venkaiah K. Indian Food Composition Tables. National Institute of Nutrition (NIN), Indian Council of Medical Research (ICMR), Hyderabad. 2017.
Source for traditional Indian dishes (dal, dosa, biryani, paneer preparations, regional breads, etc.). Provides accurate macros for home-cooked recipes that USDA does not cover.
https://www.nin.res.in
Limitations. Photo-scan and voice-log nutrition values are AI estimates of the food in front of you, not exact lab measurements. Portion sizes are approximations. Treat all AI-estimated values as guidance, not precise data.
2.Calorie & Macro Calculations
Your daily calorie target, BMR, TDEE, and macro split are calculated using these peer-reviewed equations and guidelines:
Basal Metabolic Rate (BMR) — Mifflin-St Jeor equation
Mifflin M.D., St Jeor S.T., Hill L.A., Scott B.J., Daugherty S.A., Koh Y.O. A new predictive equation for resting energy expenditure in healthy individuals. American Journal of Clinical Nutrition. 1990; 51(2): 241-247.
Currently considered the most accurate BMR equation for healthy adults aged 18+ (per Academy of Nutrition and Dietetics, 2014 recommendation). MacroLeaf AI uses Mifflin-St Jeor for every BMR calculation shown in the app.
BMR (male) = 10 × weight(kg) + 6.25 × height(cm) − 5 × age(yr) + 5
BMR (female) = 10 × weight(kg) + 6.25 × height(cm) − 5 × age(yr) − 161
https://pubmed.ncbi.nlm.nih.gov/2305711/
Total Daily Energy Expenditure (TDEE) — Activity multipliers
Food and Agriculture Organization, World Health Organization, United Nations University. Human Energy Requirements: Report of a Joint FAO/WHO/UNU Expert Consultation. FAO Food and Nutrition Technical Report Series 1. Rome. 2001 (published 2004).
TDEE = BMR × Physical Activity Level (PAL) multiplier. MacroLeaf AI uses the WHO/FAO/UNU PAL values:
Sedentary PAL = 1.2 — desk job, little/no exercise
Light PAL = 1.375 — light exercise 1-3 days/week
Moderate PAL = 1.55 — moderate exercise 3-5 days/week
Very active PAL = 1.725 — hard exercise 6-7 days/week
Athlete PAL = 1.9 — twice-daily training
https://www.fao.org/4/y5686e/y5686e00.htm
Light PAL = 1.375 — light exercise 1-3 days/week
Moderate PAL = 1.55 — moderate exercise 3-5 days/week
Very active PAL = 1.725 — hard exercise 6-7 days/week
Athlete PAL = 1.9 — twice-daily training
Macronutrient targets — IOM Dietary Reference Intakes
Institute of Medicine (US). Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids. The National Academies Press. Washington, DC. 2005.
Acceptable Macronutrient Distribution Ranges (AMDR) used as the basis for default macro splits:
Protein: 10-35% of calories (RDA 0.8 g/kg, athletes 1.2-2.0 g/kg)
Carbohydrate: 45-65% of calories
Fat: 20-35% of calories
https://nap.nationalacademies.org/catalog/10490/
Carbohydrate: 45-65% of calories
Fat: 20-35% of calories
Calorie units & conversions
USDA Agricultural Research Service. Atwater general factors used throughout the FoodData Central database.
All macro-to-calorie conversions in MacroLeaf AI use the standard Atwater factors:
1 g protein = 4 kcal
1 g carbohydrate = 4 kcal
1 g fat = 9 kcal
1 g alcohol = 7 kcal
https://www.ars.usda.gov/
1 g carbohydrate = 4 kcal
1 g fat = 9 kcal
1 g alcohol = 7 kcal
3.Weight Change Projections
Estimated weight loss/gain on the Progress tab and AI Coach replies is computed from established energy-balance principles:
Energy balance & weight change — Hall et al.
Hall K.D., Sacks G., Chandramohan D., Chow C.C., Wang Y.C., Gortmaker S.L., Swinburn B.A. Quantification of the effect of energy imbalance on bodyweight. The Lancet. 2011; 378(9793): 826-837.
Modern dynamic model of human energy balance. Replaces the older static "3500 kcal = 1 lb" rule with a model that accounts for adaptive metabolic changes during weight loss. MacroLeaf AI uses simplified deficit-to-weight projections derived from this model.
Typical rule of thumb (linear approximation):
Daily deficit ≈ 500 kcal → ~0.45 kg / 1 lb per week loss
Daily deficit ≈ 1000 kcal → ~0.9 kg / 2 lb per week loss
https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(11)60812-X/fulltext
Daily deficit ≈ 500 kcal → ~0.45 kg / 1 lb per week loss
Daily deficit ≈ 1000 kcal → ~0.9 kg / 2 lb per week loss
Individual variation is expected. Actual weight change depends on metabolic adaptation, water retention, glycogen, hormones, sleep, stress, and medication. Estimates shown in the app are statistical averages, not personalized predictions.
Doctor check. Sustained calorie deficits below 1,200 kcal/day (women) or 1,500 kcal/day (men), prolonged fasting, or rapid weight change should always be discussed with a physician — especially if you have diabetes, kidney disease, heart disease, an eating disorder, or are pregnant/breastfeeding.
4.Activity & HealthKit
Step counts, distance, and active calories displayed in the Steps card and Progress tab come from Apple HealthKit on your device — they are not transmitted to any third party.
Compendium of Physical Activities
Ainsworth B.E., Haskell W.L., Herrmann S.D., Meckes N., Bassett D.R., Tudor-Locke C., Greer J.L., Vezina J., Whitt-Glover M.C., Leon A.S. 2011 Compendium of Physical Activities: A second update of codes and MET values. Medicine & Science in Sports & Exercise. 2011; 43(8): 1575-1581.
MET (Metabolic Equivalent of Task) values used to estimate calorie burn for non-HealthKit reported activities.
https://pubmed.ncbi.nlm.nih.gov/21681120/
Apple HealthKit
Apple Inc. HealthKit framework documentation. developer.apple.com.
Step counts, distance, active energy, and floors come from Apple HealthKit and Apple's built-in motion sensors. Apple's methodology for step counting is described in their developer documentation. MacroLeaf AI reads these values; it does not modify, infer, or replace them.
https://developer.apple.com/documentation/healthkit
HealthKit data privacy. Per Apple Guideline 5.1.2(vi) and our Privacy Policy, MacroLeaf AI never transmits HealthKit data to any third party, including AI providers. Steps and active calories are used only for in-app display.
5.Hydration Targets
Default water-glass goals are based on published Adequate Intake (AI) recommendations:
Institute of Medicine — Water
Institute of Medicine (US) Panel on Dietary Reference Intakes for Electrolytes and Water. Dietary Reference Intakes for Water, Potassium, Sodium, Chloride, and Sulfate. The National Academies Press. Washington, DC. 2005.
Adequate Intake (total water, including beverages and food):
Adult men: ~3.7 L / day (~13 cups)
Adult women: ~2.7 L / day (~9 cups)
https://nap.nationalacademies.org/catalog/10925/
Adult women: ~2.7 L / day (~9 cups)
EFSA — Dietary Reference Values for Water
EFSA Panel on Dietetic Products, Nutrition, and Allergies. Scientific Opinion on Dietary Reference Values for water. EFSA Journal. 2010; 8(3): 1459.
European Adequate Intake values used to validate the IOM range for global users.
https://efsa.onlinelibrary.wiley.com/doi/10.2903/j.efsa.2010.1459
Individual needs vary. Climate, exercise, pregnancy, breastfeeding, fever, and certain medical conditions can change your water needs. The 8-glass default is a starting point, not a medical prescription.
6.AI Coach Methodology
The AI Coach is a general wellness chatbot. Its replies are bounded by safety rules and grounded in the same sources listed above. We are transparent about exactly how it works:
AI providers used
Replies are generated by one of three providers, selected automatically by latency and availability:
- Groq — Meta Llama models (primary). Privacy policy
- Google AI — Gemini models (fallback). Privacy policy
- OpenAI — GPT-4o-mini for chat, GPT-4o for vision (backup & photo scan). Privacy policy
Hardcoded medical safety rules
The AI Coach system prompt enforces these non-negotiable rules on every reply:
- Medical refusal. If the user mentions any of ~25 medical conditions (diabetes, hypertension, PCOS, kidney disease, pregnancy, cancer, etc.) the AI declines to give food advice and recommends a registered dietitian or doctor.
- No diagnosis or treatment. The AI never claims to diagnose, treat, cure, or prevent any disease.
- Supplement caution. No prescription supplement recommendations.
- Eating-disorder safeguard. Calorie targets below ~1,000 kcal/day, fasting longer than 24 hours, and purging behaviors trigger care-focused responses, not compliance.
- Pregnancy safety. No calorie deficits or weight-loss advice for pregnant or breastfeeding users.
- Photo & voice estimates. Explicitly labeled "approximate" — never presented as exact.
What the AI is grounded on
When the AI Coach gives a number or recommendation, it is based on:
- Your self-reported profile (age, gender, height, weight, activity level)
- Your daily calorie/macro targets (computed from Mifflin-St Jeor & IOM DRI)
- Today's logged meals from your Diary
- Your dietary preferences and weight goal
- The food nutrition databases above (USDA / OFF / ICMR)
It does NOT receive HealthKit data, your name, your email, your weight history, or any other personally identifying information.
AI Coach is not a doctor. All AI replies are general wellness guidance only. They are not a substitute for personalized medical advice. If you are managing a medical condition, are on prescription medication, are pregnant or breastfeeding, or have a history of disordered eating, please consult a licensed clinician before changing your diet.