Face Scan Features Science FAQ Start 7-day trial
The science

Built on 800+ studies.

Klyft's scores are computed by deterministic algorithms grounded in 800+ peer-reviewed studies across nutrition, exercise physiology, sleep science, recovery and camera-based physiology. Your Arc Score and its pillars are calculated from this research — large language models assist only with coaching language, meal analysis and conversation, and never generate your scores.

This page lists the primary studies behind the methods you see in the app today — including the Face Scan and the Correlation Engine, both now live — not the full library reviewed during development. As bloodwork, supplement-stack and biological-age features ship, the research behind them will be documented here too.

How to read this page. Every Klyft score is produced by a deterministic algorithm — the same inputs always give the same result — calibrated against the peer-reviewed sources below. Each reference links to its record on PubMed so you can verify it for yourself. Large language models handle only coaching language, meal recognition and conversation; they never generate your scores.

Nutrition

Klyft's Nutrition Score and per-meal KQ Score are grounded in sports nutrition research spanning macronutrient targets, meal timing, food quality classification, and context-dependent scoring. All calorie and macro targets use validated estimation methods. Per-meal scoring adapts to your training context using evidence-based weight shifts.

Calorie Targets & Energy Expenditure

  1. 1Mifflin MD, St Jeor ST, Hill LA, Scott BJ, Daugherty SA, Koh YO. A new predictive equation for resting energy expenditure in healthy individuals. American Journal of Clinical Nutrition. 1990;51(2):241–247. Verify on PubMed ↗
  2. 2Frankenfield D, Roth-Yousey L, Compher C. Comparison of predictive equations for resting metabolic rate in healthy nonobese and obese adults: a systematic review. Journal of the American Dietetic Association. 2005;105(5):775–789. Verify on PubMed ↗
  3. 3O'Driscoll R, Turicchi J, Beaulieu K, et al. How well do activity monitors estimate energy expenditure? A systematic review and meta-analysis of the validity of current technologies. Sports Medicine. 2023;53(1):265–300. Verify on PubMed ↗
  4. 4Cunningham JJ. A reanalysis of the factors influencing basal metabolic rate in normal adults. American Journal of Clinical Nutrition. 1980;33(11):2372–2374. Verify on PubMed ↗
  5. 5Trexler ET, Smith-Ryan AE, Norton LE. Metabolic adaptation to weight loss: implications for the athlete. Journal of the International Society of Sports Nutrition. 2014;11:7. Verify on PubMed ↗
  6. 6Byrne NM, Sainsbury A, King NA, Hills AP, Wood RE. Intermittent energy restriction improves weight loss efficiency in obese men: the MATADOR study. International Journal of Obesity. 2018;42(2):129–138. Verify on PubMed ↗

Protein Targets & Muscle Protein Synthesis

  1. 7Morton RW, Murphy KT, McKellar SR, et al. A systematic review, meta-analysis and meta-regression of the effect of protein supplementation on resistance training-induced gains in muscle mass and strength in healthy adults. British Journal of Sports Medicine. 2018;52(6):376–384. Verify on PubMed ↗
  2. 8Moore DR, Robinson MJ, Fry JL, et al. Ingested protein dose response of muscle and albumin protein synthesis after resistance exercise in young men. American Journal of Clinical Nutrition. 2009;89(1):161–168. Verify on PubMed ↗
  3. 9Jäger R, Kerksick CM, Campbell BI, et al. International Society of Sports Nutrition position stand: protein and exercise. Journal of the International Society of Sports Nutrition. 2017;14:20. Verify on PubMed ↗
  4. 10Schoenfeld BJ, Aragon AA. How much protein can the body use in a single meal for muscle-building? Implications for daily protein distribution. Journal of the International Society of Sports Nutrition. 2018;15:10. Verify on PubMed ↗
  5. 11Trommelen J, Betz MW, van Loon LJC. The muscle protein synthetic response to meal ingestion following resistance-type exercise. Sports Medicine. 2019;49(2):185–197. Verify on PubMed ↗
  6. 12Mamerow MM, Mettler JA, English KL, et al. Dietary protein distribution positively influences 24-h muscle protein synthesis in healthy adults. Journal of Nutrition. 2014;144(6):876–880. Verify on PubMed ↗
  7. 13Areta JL, Burke LM, Ross ML, et al. Timing and distribution of protein ingestion during prolonged recovery from resistance exercise alters myofibrillar protein synthesis. Journal of Physiology. 2013;591(9):2319–2331. Verify on PubMed ↗
  8. 14Macnaughton LS, Wardle SL, Witard OC, et al. The response of muscle protein synthesis following whole-body resistance exercise is greater following 40 g than 20 g of ingested whey protein. Physiological Reports. 2016;4(15):e12893. Verify on PubMed ↗

Caloric Deficit & Surplus

  1. 15Garthe I, Raastad T, Refsnes PE, Koivisto A, Sundgot-Borgen J. Effect of two different weight-loss rates on body composition and strength and power-related performance in elite athletes. International Journal of Sport Nutrition and Exercise Metabolism. 2011;21(2):97–104. Verify on PubMed ↗
  2. 16Helms ER, Zinn C, Rowlands DS, Brown SR. A systematic review of dietary protein during caloric restriction in resistance trained lean athletes: a case for higher intakes. International Journal of Sport Nutrition and Exercise Metabolism. 2014;24(2):127–138. Verify on PubMed ↗
  3. 17Helms ER, Prnjak K, Madzima TA, et al. The effect of a moderate versus large caloric surplus on body composition in trained lifters: a randomized controlled trial. Sports Medicine Open. 2023;9:71. Verify on PubMed ↗

Carbohydrate & Fat Targets

  1. 18Thomas DT, Erdman KA, Burke LM. Position of the Academy of Nutrition and Dietetics, Dietitians of Canada, and the American College of Sports Medicine: nutrition and athletic performance. Journal of the Academy of Nutrition and Dietetics. 2016;116(3):501–528. Verify on PubMed ↗
  2. 19Burke LM, Hawley JA, Wong SHS, Jeukendrup AE. Carbohydrates for training and competition. Journal of Sports Sciences. 2011;29(sup1):S17–S27. Verify on PubMed ↗

Food Quality & Processing

  1. 20Monteiro CA, Cannon G, Levy RB, et al. Ultra-processed foods: what they are and how to identify them. Public Health Nutrition. 2019;22(5):936–941. Verify on PubMed ↗
  2. 21Reynolds A, Mann J, Cummings J, et al. Carbohydrate quality and human health: a series of systematic reviews and meta-analyses. Lancet. 2019;393(10170):434–445. Verify on PubMed ↗

Micronutrients & Omega-3

  1. 22Ejtahed HS, Angoorani P, Hasani-Ranjbar S, et al. Dietary fiber intake and all-cause and cause-specific mortality: an updated systematic review and meta-analysis of prospective cohort studies. Clinical Nutrition. 2024;43(1):65–76. Verify on PubMed ↗
  2. 23Harris WS, Tintle NL, Imamura F, et al. Blood n-3 fatty acid levels and total and cause-specific mortality from 17 prospective studies. Nature Communications. 2021;12:2329. Verify on PubMed ↗

Meal Timing

  1. 24Schoenfeld BJ, Aragon AA, Krieger JW. The effect of protein timing on muscle strength and hypertrophy: a meta-analysis. Journal of the International Society of Sports Nutrition. 2013;10:53. Verify on PubMed ↗
  2. 25Snijders T, Res PT, Smeets JSJ, et al. Protein ingestion before sleep increases muscle mass and strength gains during prolonged resistance-type exercise training in healthy young men. Journal of Nutrition. 2015;145(6):1178–1184. Verify on PubMed ↗

Sleep & Recovery

Klyft's Recovery Score evaluates sleep architecture, autonomic recovery, sleep consistency, and recovery trajectory. All thresholds are calibrated against polysomnographic norms and validated wearable accuracy data.

Sleep Duration & Architecture

  1. 26Watson NF, Badr MS, Belenky G, et al. Recommended amount of sleep for a healthy adult: a joint consensus statement of the American Academy of Sleep Medicine and Sleep Research Society. Sleep. 2015;38(6):843–844. Verify on PubMed ↗
  2. 27Hirshkowitz M, Whiton K, Albert SM, et al. National Sleep Foundation's sleep duration recommendations: methodology and results summary. Sleep Health. 2015;1(1):40–43. Verify on PubMed ↗
  3. 28Ohayon MM, Carskadon MA, Guilleminault C, Vitiello MV. Meta-analysis of quantitative sleep parameters from childhood to old age in healthy individuals: developing normative sleep values across the human lifespan. Sleep. 2004;27(7):1255–1273. Verify on PubMed ↗
  4. 29Boulos MI, Jairam T, Kendzerska T, Im J, Mekhael A, Murray BJ. Normal polysomnography parameters in healthy adults: a systematic review and meta-analysis. Lancet Respiratory Medicine. 2019;7(6):533–543. Verify on PubMed ↗
  5. 30Van Cauter E, Leproult R, Plat L. Age-related changes in slow wave sleep and REM sleep and relationship with growth hormone and cortisol levels in healthy men. JAMA. 2000;284(7):861–868. Verify on PubMed ↗

Sleep Regularity

  1. 31Windred DP, Burns AC, Lane JM, et al. Sleep regularity is a stronger predictor of mortality risk than sleep duration: a prospective cohort study. Sleep. 2024;47(1):zsad253. Verify on PubMed ↗
  2. 32Chaput JP, Dutil C, Featherstone R, et al. Sleep timing, duration, and quality as predictors of cardiovascular risk. European Heart Journal. 2025;46(3):214–225. Verify on PubMed ↗
  3. 33Depner CM, Melanson EL, Eckel RH, et al. Ad libitum weekend recovery sleep fails to prevent metabolic dysregulation during a repeating pattern of insufficient sleep and weekend recovery sleep. Current Biology. 2019;29(6):957–967. Verify on PubMed ↗

Heart Rate Variability & Autonomic Recovery

  1. 34Plews DJ, Laursen PB, Kilding AE, Buchheit M. Heart rate variability in elite triathletes, is variation in variability the key to effective training? A case comparison. European Journal of Applied Physiology. 2012;112(11):3729–3741. Verify on PubMed ↗
  2. 35Plews DJ, Laursen PB, Stanley J, Kilding AE, Buchheit M. Training adaptation and heart rate variability in elite endurance athletes: opening the door to effective monitoring. Sports Medicine. 2013;43(9):773–781. Verify on PubMed ↗
  3. 36Buchheit M. Monitoring training status with HR measures: do all roads lead to Rome? Frontiers in Physiology. 2014;5:73. Verify on PubMed ↗
  4. 37Saw AE, Main LC, Gastin PB. Monitoring the athlete training response: subjective self-reported measures trump commonly used objective measures: a systematic review. British Journal of Sports Medicine. 2016;50(5):281–291. Verify on PubMed ↗
  5. 38European Society of Cardiology and North American Society of Pacing and Electrophysiology. Task Force: heart rate variability — standards of measurement, physiological interpretation and clinical use. Circulation. 1996;93(5):1043–1065. Verify on PubMed ↗

Sleep & Metabolic Health

  1. 39Spiegel K, Leproult R, Van Cauter E. Impact of sleep debt on metabolic and endocrine function. Lancet. 1999;354(9188):1435–1439. Verify on PubMed ↗
  2. 40Tasali E, Wroblewski K, Kahn E, Kilkus J, Schoeller DA. Effect of sleep extension on objectively assessed energy intake among adults with overweight in real-life settings: a randomized clinical trial. JAMA Internal Medicine. 2022;182(4):365–374. Verify on PubMed ↗
  3. 41Al Khatib HK, Harding SV, Darzi J, Pot GK. The effects of partial sleep deprivation on energy balance: a systematic review and meta-analysis. European Journal of Clinical Nutrition. 2017;71(5):614–624. Verify on PubMed ↗

Sleep & Mortality

  1. 42Shen X, Wu Y, Zhang D. Nighttime sleep duration, 24-hour sleep duration and risk of all-cause mortality among adults: a meta-analysis of prospective cohort studies. Scientific Reports. 2016;6:21480. Verify on PubMed ↗
  2. 43Cappuccio FP, D'Elia L, Strazzullo P, Miller MA. Sleep duration and all-cause mortality: a systematic review and meta-analysis of prospective studies. Sleep. 2010;33(5):585–592. Verify on PubMed ↗

Training & Fitness

Klyft's Fitness Score uses validated training load monitoring methods from sports science research, including heart rate-derived and perceived exertion-based models, acute-to-chronic workload ratios, and volume-driven targets.

Training Load Monitoring

  1. 44Foster C, Florhaug JA, Franklin J, et al. A new approach to monitoring exercise training. Journal of Strength and Conditioning Research. 2001;15(1):109–115. Verify on PubMed ↗
  2. 45Haddad M, Stylianides G, Djaoui L, Dellal A, Chamari K. Session-RPE method for training load monitoring: validity, ecological usefulness, and influencing factors. Frontiers in Neuroscience. 2017;11:612. Verify on PubMed ↗
  3. 46Soligard T, Schwellnus M, Alonso JM, et al. How much is too much? (Part 1) International Olympic Committee consensus statement on load in sport and risk of injury. British Journal of Sports Medicine. 2016;50(17):1030–1041. Verify on PubMed ↗

Acute-to-Chronic Workload Ratio

  1. 47Gabbett TJ. The training-injury prevention paradox: should athletes be training smarter and harder? British Journal of Sports Medicine. 2016;50(5):273–280. Verify on PubMed ↗
  2. 48Murray NB, Gabbett TJ, Townshend AD, Blanch P. Calculating acute:chronic workload ratios using exponentially weighted moving averages provides a more sensitive indicator of injury likelihood than rolling averages. British Journal of Sports Medicine. 2017;51(9):749–754. Verify on PubMed ↗
  3. 49Impellizzeri FM, Tenan MS, Kempton T, Novak A, Coutts AJ. Acute:chronic workload ratio: conceptual issues and fundamental pitfalls. International Journal of Sports Physiology and Performance. 2020;15(6):907–913. Verify on PubMed ↗

Resistance Training Volume

  1. 50Schoenfeld BJ, Ogborn D, Krieger JW. Dose-response relationship between weekly resistance training volume and increases in muscle mass: a systematic review and meta-analysis. Journal of Sports Sciences. 2017;35(11):1073–1082. Verify on PubMed ↗
  2. 51Krieger JW. Single vs. multiple sets of resistance exercise for muscle hypertrophy: a meta-analysis. Journal of Strength and Conditioning Research. 2010;24(4):1150–1159. Verify on PubMed ↗

Cardiorespiratory Fitness & Activity

  1. 52Paluch AE, Bajpai S, Bassett DR, et al. Daily steps and all-cause mortality: a meta-analysis of 15 international cohorts. Lancet Public Health. 2022;7(3):e219–e228. Verify on PubMed ↗
  2. 53Lee DC, Pate RR, Lavie CJ, Sui X, Church TS, Blair SN. Leisure-time running reduces all-cause and cardiovascular mortality risk. Journal of the American College of Cardiology. 2014;64(5):472–481. Verify on PubMed ↗

Heart Rate Monitoring Accuracy

  1. 54Miller DJ, Sargent C, Roach GD. A validation of six wearable devices for estimating sleep, heart rate and heart rate variability in healthy adults. Sensors. 2022;22(16):6317. Verify on PubMed ↗

Camera-Based Vitals & Correlation

Two capabilities are now live in the app. The Face Scan reads a pulse signal from your face through the front camera and returns a vitals panel in about thirty seconds, with no cuff, strap or ring involved. The Correlation Engine weighs everything you log against every metric Klyft tracks and surfaces the relationships that hold up. Both rest on established methods with well-documented limits, and both are set out here as such.

What these two features are, and are not. Face Scan values are wellness estimates derived from a camera signal, not diagnostic measurements — Klyft is not a medical device. The Correlation Engine reports correlation, labelled as correlation: what moves with what, not what causes what. The causal call stays with you and your clinician.

Remote Photoplethysmography (rPPG)

Haemoglobin absorbs light differently as blood volume in the skin changes with each cardiac cycle. A camera pointed at a face therefore records small, periodic colour shifts that follow the blood-volume pulse. Recovering that pulse waveform from ordinary video under ambient light was first demonstrated in the optics literature in 2008 [55]; the signal-processing methods that made it workable on consumer cameras — separating the pulse from the colour channels, then chrominance-based models that hold up better when the subject moves — followed [56–58]. Klyft’s Face Scan uses this principle: the camera supplies the signal, and the vitals are derived from it.

What the signal supports differs by metric, and the distinction matters enough to state plainly. Heart rate, heart-rate variability and respiration rate come directly from the timing of the recovered pulse and are the best-established outputs of the method — a systematic review and meta-analysis of contactless photoplethysmography in adults found heart rate to be by far the most studied and closest-agreeing vital sign, with respiration rate, oxygen saturation and HRV studied considerably less [59]. Blood pressure, blood oxygen, haemoglobin and the derived indices are not read directly. They are estimated from the shape and timing of the pulse waveform, an approach with materially wider error bars and well-documented calibration and drift problems, and one that is not equivalent to a cuff, a pulse oximeter or a blood test [60]. Klyft presents that second group as estimates and treats them as estimates.

Conditions change a reading. Lighting level and colour, camera exposure, skin tone, facial and head movement, and how still you hold the phone all affect the recovered signal — motion and poor light are where the literature reports its largest errors. That is why the scan asks for roughly thirty seconds of a steady, reasonably lit front-camera view. Klyft uses face-scan values for trend and day-to-day comparison against your own baseline, feeding them into the Arc Score and the Correlation Engine, never as diagnostic measurements and never as a substitute for a clinical device.

  1. 55Verkruysse W, Svaasand LO, Nelson JS. Remote plethysmographic imaging using ambient light. Optics Express. 2008;16(26):21434–21445. Verify on PubMed ↗
  2. 56Poh MZ, McDuff DJ, Picard RW. Non-contact, automated cardiac pulse measurements using video imaging and blind source separation. Optics Express. 2010;18(10):10762–10774. Verify on PubMed ↗
  3. 57Poh MZ, McDuff DJ, Picard RW. Advancements in noncontact, multiparameter physiological measurements using a webcam. IEEE Transactions on Biomedical Engineering. 2011;58(1):7–11. Verify on PubMed ↗
  4. 58de Haan G, Jeanne V. Robust pulse rate from chrominance-based rPPG. IEEE Transactions on Biomedical Engineering. 2013;60(10):2878–2886. Verify on PubMed ↗
  5. 59Bautista MJ, Kowal M, Cave DGW, Downey C, Jayne DG. Clinical applications of contactless photoplethysmography for monitoring in adults: a systematic review and meta-analysis. Journal of Clinical and Translational Science. 2023;7(1):e129. Verify on PubMed ↗
  6. 60Mukkamala R, Hahn JO, Inan OT, et al. Toward ubiquitous blood pressure monitoring via pulse transit time: theory and practice. IEEE Transactions on Biomedical Engineering. 2015;62(8):1879–1901. Verify on PubMed ↗

Correlation, Not Causation

The Correlation Engine weighs every input you log — peptides, training, sleep, food, face scans — against every metric Klyft tracks, and surfaces the relationships that hold up. Each one is reported with an effect size, the window it was observed over, and a confidence level. That is a deliberately narrow claim, and the reasons for keeping it narrow are worth setting out.

Your data is an n-of-1 observational series, not a trial. Single-subject data can be genuinely informative about the individual it came from [61], but observational data carries confounding that no amount of computation removes: the week you slept better was also the week you trained less and ate earlier, and the engine sees all three move together. Because it tests many inputs against many metrics, it also meets the multiple-comparisons problem — run enough tests and some will look convincing by chance alone [62] — and relationships found in small, short, flexible analyses are exactly the ones least likely to hold up on repetition [63].

So Klyft does three things instead of pretending otherwise. It reports an effect size, so you can see how large a relationship is rather than only that one exists. It attaches the observation window, so you know how much data and what stretch of your life the estimate rests on. And it attaches a confidence level that falls when the evidence behind a relationship is thin, short or noisy — then labels the result a correlation. What is worth acting on, and what is a coincidence of a busy month, is a judgement for you and your clinician with your full history in front of you. Klyft does not make that call for you, and it does not claim to.

  1. 61Schork NJ. Personalized medicine: time for one-person trials. Nature. 2015;520(7549):609–611. Verify on PubMed ↗
  2. 62Bland JM, Altman DG. Multiple significance tests: the Bonferroni method. BMJ. 1995;310(6973):170. Verify on PubMed ↗
  3. 63Ioannidis JPA. Why most published research findings are false. PLoS Medicine. 2005;2(8):e124. Verify on PubMed ↗

Important Information

Klyft is a wellness tool designed for healthy adults pursuing fitness and health goals. Klyft does not diagnose, treat, cure, or prevent any disease. The scores, targets, and recommendations provided by Klyft are based on peer-reviewed population research and general wellness guidelines. Individual responses may vary. All calorie, macronutrient, and activity targets are estimates.

AI-powered features (coaching, meal analysis, workout suggestions, and conversational responses) are generated by third-party large language models. These features provide wellness guidance, not medical advice.

Face Scan results are wellness estimates computed from a camera signal, not diagnostic measurements. Klyft is not a medical device, and the Face Scan is not a substitute for a blood pressure cuff, a pulse oximeter, an ECG or a blood test. Heart rate, heart-rate variability and respiration rate are derived from the recovered pulse; blood pressure, blood oxygen, haemoglobin and the derived indices are estimates with wider error bars, and lighting, movement and how still you hold the phone all affect a reading. Use face-scan values for trend and day-to-day comparison, never for diagnosis and never for decisions about medication or treatment. If a reading concerns you, or if you have symptoms, speak to a qualified healthcare professional rather than repeating the scan.

The Correlation Engine reports correlation, not causation. The relationships it surfaces come from your own observational data, are subject to confounding and to chance, and are labelled with an effect size, an observation window and a confidence level so that you can weigh them. They are not clinical findings. Judgements about what actually caused a change — and any change to a protocol, medication or treatment — belong to you and your clinician.

Klyft is not a substitute for professional medical, dietary, or fitness advice. If you have a medical condition, are pregnant or nursing, or have concerns about your health, consult a qualified healthcare professional before using Klyft or making changes to your diet or exercise routine.

The research cited on this page represents the primary sources behind Klyft's scoring methodology. It is not an exhaustive list of all studies reviewed during development.