Supplements for Wearable Metrics: What Works, What's Hype
- Ryan - Kygo Health
- Jun 16
- 10 min read
Updated: 1 day ago
Last Updated: June 16, 2026
Of the 27 supplements studied for the six metrics your wearable tracks (sleep latency, deep sleep, staying asleep, HRV, resting heart rate, and recovery), only a short list has strong research behind it. Omega-3 lowers resting heart rate and raises HRV. Glycine and magnesium help you fall asleep faster. Ashwagandha helps you stay asleep. The rest are weaker, mixed, or were tested and found to do nothing. This post ranks every supplement by metric and evidence grade, flags which trials were funded by the company selling the product, and shows what the research actually supports.
That last part matters more than any single supplement, because most of the popular advice skips it.
You see the number every morning. The supplement aisle has an answer for all of it.
Open your ring or watch and there it is: an HRV figure, a deep sleep total, a readiness score. You half understand it, and within a day or two an ad promises a capsule that will fix it. Magnesium for sleep. Ashwagandha for stress. Some powder for "recovery."
The problem is not that supplements never work. It is that the marketing flattens a messy evidence base into a single confident claim. A figure like "27 randomized trials" turns out to be three small studies. A "deep sleep" benefit turns out to have been measured in mice. A glowing recovery result turns out to have been funded by the brand that sells the capsule.
So we did the boring version. We took the six metrics a wearable reports, pulled every supplement with metric specific research, and graded it. Strong means a meta analysis or multiple randomized controlled trials. Moderate means one solid trial or several consistent smaller ones. Weak means a single small, pilot, or preclinical study. And we kept the nulls, the supplements that were tested and showed no real effect, because those save you the most money.
The six metrics, and what actually moves each one
Each metric below lists the supplements with real evidence first, then the weaker or mixed ones, and the funding flag where it applies. "Benefit" means the direction your wearable rewards: faster onset, more deep sleep, less waking, higher HRV, lower resting heart rate, higher recovery.
Falling asleep faster (sleep latency)
Supplement | Effect | Evidence | Flag |
Melatonin | Onset about 7 min faster; peaks near 4 mg (Ferracioli-Oda 2013 meta, n=1,683) | Strong | |
Ashwagandha | Onset shorter, SMD -0.53; 600 mg best (Cheah 2021 meta, n=400) | Strong | Industry funded |
Magnesium | Onset about 17 min faster (Mah and Pitre 2021 meta, n=151 older adults) | Moderate | Low trial quality |
Glycine (3g) | Faster onset and faster slow wave sleep (Yamadera 2007, n=11) | Moderate | |
GABA | Onset 13.4 to 5.7 min (Byun 2018, n=40) | Moderate | |
L-theanine | Subjective only, no objective change (Bulman 2025 meta, n=897) | Weak | Subjective |
Tested with no real effect: valerian, 5-HTP, tart cherry, CBD isolate.
Deep sleep (slow wave sleep)
Supplement | Effect | Evidence | Flag |
Glycine (3g) | Faster deep sleep onset (Yamadera 2007, n=11) | Moderate | |
Magnesium L-threonate | More deep sleep and readiness on Oura (Hausenblas/Breus 2024) | Moderate | All manufacturer funded |
Magnesium (oral) | About 6 more min of slow wave sleep in older adults (Held 2002, n=12) | Weak | Small, elderly |
Tart cherry | More total sleep time (Losso 2018, n=8) | Weak | Very small |
Melatonin | Boosts NREM and delta power via MT2 (Comai 2024, mechanism) | Moderate | Mainly a timing aid |
The myth here: L-theanine and ashwagandha "deep sleep" claims come from animal and in vitro models, not human sleep studies.
Staying asleep (less waking)
Supplement | Effect | Evidence | Flag |
Ashwagandha (600 mg) | Less waking, SMD -0.39 (Cheah 2021 meta) | Strong | Industry funded |
Tart cherry | About 17 min less waking (Pigeon 2010, n=15) | Moderate | Small |
Glycine (3g) | Reduced waking (Yamadera 2007, n=11) | Moderate | |
Melatonin (standard) | No significant effect on waking (Moon 2022 meta) | Strong null |
The most bought form of melatonin, the standard immediate release kind, does not help you stay asleep. It helps you fall asleep and shift your clock, not maintain sleep.
HRV (higher RMSSD or HF power)
Supplement | Effect | Evidence | Flag |
Omega-3 (EPA/DHA) | Consistently raises HF power; most studied dietary HRV factor (multiple meta analyses) | Strong | |
Ashwagandha (Witholytin) | RMSSD rose vs placebo decline (Smith/Lopresti 2023, n=111) | Strong | Single proprietary extract trial |
GABA | Raises HRV via parasympathetic shift (Guimaraes 2024, n=30) | Moderate | All female plus exercise |
Magnesium L-threonate | RMSSD up about 1.45 ms (Lopresti 2025, Oura, n=100) | Moderate | Manufacturer funded |
Vitamin D, B12, multivitamin | Help mainly when you are deficient (micronutrient review 2022) | Moderate | Deficiency correction, not a booster |
The myth here: L-theanine blunts the stress response acutely, but that is not the same as a higher resting HRV.
Resting heart rate (lower is the benefit)
Supplement | Effect | Evidence | Flag |
Omega-3 (DHA) | Lower by about 2.23 bpm overall; DHA does the work, EPA does not (Hidayat 2018 meta, ~3,000 people) | Strong | |
Chromium, oat bran fiber | Small reductions in specific patient groups only | Weak | Narrow population |
Tested with no real effect on resting heart rate: potassium (Gijsbers 2016 meta), vitamin D, L-arginine, nitrate/beetroot, creatine, CoQ10. Omega-3 is the one supplement with strong evidence here, and even then the effect is small.
Recovery and readiness score
Recovery and readiness scores are proprietary blends of HRV, resting heart rate, and sleep measured against your own baseline. There is no clinical gold standard for "recovery," so treat these as best available, not validated.
Supplement | Effect | Evidence | Flag |
Omega-3 (DHA) | Lowers resting heart rate, a core input | Moderate | |
Saffron | Better sleep quality, PSQI -2.14 (Sleep Medicine 2022 meta) | Moderate | |
Glycine (3g) | Faster onset, better quality, lower core temp | Moderate | Small samples |
Tart cherry | More sleep, raises melatonin, eases soreness | Moderate | Small pilots |
Ashwagandha | Lowers cortisol, improves sleep, some HRV gain | Moderate | Proprietary extract trials |
Do not rely on these for recovery: CBD (no meaningful change on WHOOP), GABA (poor oral absorption), creatine (a cognitive buffer when underslept, not a recovery aid), apigenin (no human sleep trial exists).
Want this as a tool you can filter instead of a wall of tables?
We built the interactive Supplements by Metric explorer. Pick the metric you care about, and it groups every supplement into what is backed, what is weaker or mixed, and what was tested with no effect. Each card opens to the finding, the study, the evidence grade, the funding flag, and a link to the source.
The reason any of this matters for you specifically is simple: a meta analysis tells you what happens on average across hundreds of people. It cannot tell you what magnesium does to your sleep latency or what omega-3 does to your morning HRV. That is the gap Kygo is built to close, by lining up what you actually take and eat against the numbers your wearable already records. Download Kygo on iOS or Android and see whether the research pattern shows up in your own data.
The full matrix: every supplement against every metric
S = Strong evidence
M = Moderate evidence
W = Weak evidence
A Check with no grade = A measured benefit
"no effect" = It was tested and did nothing for that metric.
A flag = At least one underlying trial was industry funded or otherwise limited.
Supplement | Sleep latency | Deep sleep | Staying asleep | HRV | RHR | Recovery |
Omega-3 (EPA/DHA) | W | S | S | M | ||
Ashwagandha (flag) | S | W | S | S | no effect | M |
Glycine (3g) | M | M | M | M | ||
Magnesium (oral) | M | W | W | mixed | no effect | W |
Magnesium L-threonate (flag) | M | M | M | |||
Melatonin | S | M | no effect | mixed | W | |
Tart cherry | no effect | W | M | M | ||
GABA (flag) | M | M | no effect | |||
L-theanine | M (subjective) | W | mixed | M (acute) | no effect | M (acute) |
Saffron | M | |||||
Chamomile (whole extract) | W | |||||
Nitrate/beetroot | acute | no effect | M (acute) | |||
Tryptophan (1g+) | W | |||||
5-HTP | W | |||||
Valerian | no effect | no effect | no effect | |||
CBD (flag) | no effect | mixed | no effect | |||
Vitamin D | no effect | M (deficiency) | no effect | |||
Vitamin B12 | M (deficiency) | |||||
Zinc | W | |||||
Multivitamin | M | |||||
Potassium | no effect | |||||
L-arginine | no effect | |||||
Chromium (flag) | W | |||||
Oat bran fiber | W | |||||
Creatine (flag) | no effect | |||||
Passionflower (flag) | W (subjective) | |||||
PeptiSleep (flag) | W | W | W |
The honesty layer: the claims that fall apart on a second look
This is the part the ads leave out. Each of these is a popular claim that does not survive a look at the primary research.
The magnesium "27 randomized trials" figure is really three small studies, about 151 people total, all in older adults. Magnesium helping you fall asleep faster is real and worth knowing. The scale of the claim is not.
Standard melatonin does nothing for staying asleep. A meta analysis found no significant effect on waking for the immediate release form, which is the kind most people buy. It is a falling asleep and clock shifting tool, not a sleep maintenance tool.
CBD shows no meaningful recovery change on WHOOP, and CBD isolate did nothing for sleep onset in a controlled trial. The category is enormous and the metric specific evidence is close to empty.
Apigenin, the trendy "the active part of chamomile" capsule, has zero human sleep trials. The chamomile evidence that exists is for the whole extract and is about sleep quality, not insomnia.
Passionflower's "evidence" was self reported, never measured on a sleep study, and funded by a botanicals company.
L-theanine calms the acute stress response, which is genuinely useful before a stressful moment. It is not the same as a higher resting HRV or faster objective sleep onset, and the data does not support stretching it that far.
The part no supplement post tells you: your wearable may not even see it
Two cautions before you spend a cent.
First, funding. Nearly every supplement specific trial in this set was paid for by the company selling the supplement. That does not automatically make a result wrong, but it is the single biggest reason to read effect sizes skeptically and to want independent replication. We flag it because saying so is the honest move.
Second, detectability. Look at the effect sizes. Resting heart rate down about 2.2 bpm. HRV up about 1.45 ms. Those are real averages in studies, and they are also small relative to how much your numbers bounce around night to night on their own. Many of these results were measured on Oura specifically. So even a supplement with strong evidence may produce a change your ring or watch cannot cleanly separate from normal noise. For more on how much your device can and cannot detect, see our breakdowns of HRV factors ranked by evidence, how to fall asleep faster, and how to stay asleep.
This is why "research shows X affects this metric" is the only honest frame, and "take X to fix your number" is not. The first is what the studies support. The second is what the ads sell.
The only way to know if it works for you
Averages are a starting point, not an answer. The supplement that moved HRV in a trial of 100 people may do nothing for you, and the one the research shrugged at might line up with your best mornings. The only way to tell is to watch your own numbers while you change one thing at a time.
That is the whole idea behind Kygo. It connects what you take and eat to the biometrics your wearable already records across Oura, Apple Health, Garmin, Fitbit, and WHOOP, then surfaces the patterns over a 12 to 36 hour window so you can see whether a supplement shows up in your data or not.
Start with the research, then check it against yourself. Download Kygo on iOS or Android, or learn more at www.kygo.app.
Frequently asked questions
What supplements actually have strong evidence for wearable metrics?
Across the six metrics, the strongest evidence belongs to omega-3 (lower resting heart rate, higher HRV), ashwagandha (faster sleep onset, less waking, higher HRV), glycine (faster onset and deep sleep), magnesium (faster onset), saffron (sleep quality), and tart cherry (less waking). Most have at least one industry funded study behind them, so read the effect sizes with that in mind.
Does magnesium really help sleep?
Magnesium has moderate evidence for helping you fall asleep faster, about 17 minutes in a meta analysis. But that meta analysis was three small trials totaling roughly 151 older adults, not the "27 trials" figure you often see quoted. It is a real but modest effect.
Does melatonin help you stay asleep?
No. A meta analysis found standard immediate release melatonin, the most common form sold, has no significant effect on waking during the night. It helps you fall asleep faster and shift your sleep timing, not stay asleep.
What is the best supplement for HRV?
Omega-3 has the strongest and most studied evidence for raising HRV, followed by ashwagandha in a proprietary extract trial. Vitamin D, B12, and a multivitamin mainly help if you are deficient. Keep in mind the changes are small relative to night to night variation.
Does CBD improve recovery or sleep?
The metric specific evidence is weak. CBD showed no meaningful recovery change on WHOOP, and CBD isolate did nothing for sleep onset in a controlled trial. It is one of the most marketed and least supported categories for these specific numbers.
Why do so many supplement studies disagree?
Small sample sizes, different doses and forms, subjective versus objective measurement, and heavy industry funding all push results around. A subjective sleep survey can show a benefit that a sleep lab measurement does not, which is why we separate the two.
Will my wearable actually detect a supplement's effect?
Often not cleanly. Many proven effects are small, for example resting heart rate down around 2 bpm, which can be smaller than your normal day to day swing. The most reliable approach is to track one change over many days and look for a pattern rather than judging a single night.
How can I test a supplement on my own data?
Pick one supplement, keep everything else as steady as you can, and watch the relevant metric for two to three weeks. Tools like Kygo line up what you take against your wearable data so you can see whether the research pattern appears in your own numbers instead of guessing.
Sources
Ferracioli-Oda E, et al. Meta-analysis of melatonin for sleep (2013), n=1,683.
Cheah KL, et al. Ashwagandha and sleep, systematic review and meta-analysis (2021), n=400.
Mah J, Pitre T. Oral magnesium and sleep, meta-analysis (2021), n=151.
Yamadera W, et al. Glycine and sleep quality, PSG study (2007), n=11.
Byun JI, et al. GABA and sleep latency, PSG (2018), n=40.
Bulman A, et al. L-theanine and sleep, meta-analysis (2025), n=897.
Held K, et al. Magnesium and slow wave sleep in the elderly (2002), n=12.
Hausenblas HA, Breus M, et al. Magnesium L-threonate and Oura sleep metrics (2024).
Lopresti AL, et al. Magnesium L-threonate, HRV and RHR on Oura (2025), n=100.
Losso JN, et al. Tart cherry and total sleep time (2018), n=8.
Comai S, et al. Melatonin MT2 receptor and NREM delta power (2024), J Pineal Research.
Pigeon WR, et al. Tart cherry and WASO in insomnia (2010), n=15.
Moon E, et al. Melatonin and sleep maintenance, meta-analysis (2022), Neuropsychopharmacology.
Shinjyo N, et al. Valerian for sleep, meta-analysis (2020), n=6,894.
Hidayat K, et al. Omega-3 and resting heart rate, meta-analysis (2018), 51 RCTs.
Smith/Lopresti, Witholytin ashwagandha and HRV, RCT (2023), n=111.
Gijsbers L, et al. Potassium and blood pressure and heart rate, meta-analysis (2016), n=1,086.
Saffron and sleep quality, meta-analysis, Sleep Medicine (2022), PSQI -2.14.
Micronutrients and HRV review, Advances in Nutrition (2022), vitamin D, B12, multivitamin, zinc.
Disclaimer: Kygo is a personal data aggregation and insights platform designed for informational purposes only. The information provided by Kygo, including correlations, patterns, and trends identified in your data, does not constitute medical advice, diagnosis, or treatment. Always consult a licensed healthcare provider with any questions regarding medical conditions or before starting any supplement.
Which supplement claim surprised you most, the one that held up or the one that fell apart? Tell us what you have tracked.