NEUROFIT Vitals SDK
Chest-Strap-Grade HRV
From The Phone Camera Your Users Already Own
On-device heart rate, HRV (RMSSD) and breathing rate from a 60-second fingertip reading on iOS and Android. Validated against a Polar H10 chest strap across 30 participants, 265 readings and 20 phone models.
0.47 bpm
Heart rate error vs. chest strap
4.0 ms
RMSSD error vs. chest strap
0 of 265
Readings off by 20 ms or more
r = 0.975
RMSSD correlation with strap
What The SDK Delivers
The same engine that powers Vitals readings in the NEUROFIT app, packaged for your product. Your user covers the rear camera and flash with a fingertip for 60 seconds; the SDK returns the metrics, a quality tier and diagnostics.
Heart rate, RMSSD, SDNN, Baevsky stress index, Poincaré SD2:SD1 and breathing rate from one 60-second fingertip reading
Runs entirely on the device using deterministic signal processing: no machine learning, no images stored or transmitted, and only the computed metrics returned to your app
Native Swift and Kotlin engines with parity-tested outputs, so iOS and Android report the same numbers for the same signal
Live contact and signal-quality guidance during the reading, plus a per-reading quality tier
A consistency gate that withholds an unreliable reading instead of showing a wrong number
Validated Against A Chest Strap
Thirty volunteers each took ten 60-second readings with the app, five at rest and five during paced breathing at 6 breaths per minute, while wearing a Polar H10 chest strap recording every heartbeat. Readings the SDK accepted were compared with the strap beat for beat. The figures below are for those accepted readings; the SDK withheld 3.3% of readings as unreliable.
| Metric | Mean abs. error | r | 95% limits of agreement | Notes |
|---|---|---|---|---|
| Heart rate | 0.47 bpm | 0.999 | −1.4 to +0.9 bpm | 90% of readings within 1 bpm |
| RMSSD | 4.03 ms | 0.975 | −9.2 to +11.7 ms | 92.5% within 10 ms; none off by 20 ms or more |
| SDNN | 4.13 ms | 0.984 | −9.8 to +12.0 ms | |
| Baevsky stress index (ln) | 0.18 | 0.970 | −0.51 to +0.42 | 78% of readings within ±25% of the strap value |
| Poincaré SD2:SD1 | 0.36 | 0.852 | −1.05 to +0.94 | Error tracks RMSSD error |
| Breathing rate (paced 6 bpm) | 0.46 bpm | n/a | n/a | 89% within 1 bpm; confidence-gated |
Left: camera RMSSD against the strap for every accepted reading. Right: Bland-Altman plot with the 95% limits of agreement; dotted lines mark ±20 ms.
Consistent Across Skin Tones, Genders And Ages
RMSSD error for the accepted readings in each group. No group had a reading off by 20 ms or more. Groups with one or two participants describe those individuals rather than the group.
| Group | Participants | Readings | RMSSD error (ms) | Within 10 ms |
|---|---|---|---|---|
| Platform | ||||
| iOS | 21 | 178 | 3.76 | 93% |
| Android | 9 | 87 | 4.58 | 91% |
| Condition | ||||
| Resting | 30 | 134 | 3.79 | 94% |
| Paced breathing | 30 | 131 | 4.27 | 91% |
| Monk skin tone | ||||
| Monk 2 | 4 | 33 | 2.83 | 97% |
| Monk 3 | 3 | 28 | 1.60 | 100% |
| Monk 4 | 5 | 46 | 4.56 | 89% |
| Monk 5 | 9 | 79 | 4.23 | 89% |
| Monk 6 | 6 | 53 | 5.25 | 92% |
| Monk 8 | 3 | 26 | 4.10 | 96% |
| Gender | ||||
| Female | 12 | 105 | 4.31 | 90% |
| Male | 18 | 160 | 3.85 | 94% |
| Age range | ||||
| 18–24 | 1 | 8 | 2.07 | 100% |
| 25–34 | 21 | 181 | 4.03 | 94% |
| 35–44 | 5 | 47 | 4.77 | 89% |
| 45–54 | 1 | 10 | 1.28 | 100% |
| 55–64 | 2 | 19 | 4.48 | 79% |
How The Study Was Run
Reference
Polar H10 beat-to-beat intervals, cleaned of isolated premature beats before computing the reference values. Strap records with 5% or more ectopic beats were excluded, since HRV is defined on sinus rhythm; this removed 3 readings.
Frozen before testing
Every constant in the estimator and the consistency gate was frozen before the participants used to test it were enrolled, and the two are reported separately. The iOS pipeline has been tested on an independent holdout of 5 participants and 40 accepted readings: RMSSD error 3.13 ms, none off by 20 ms or more. The Android-specific components were finalized after the first two Android holdout participants and are being tested on new participants now.
What we don't claim yet
The sample has no participant at Monk skin tones 7, 9 or 10 and is concentrated at ages 25–34; recruitment is ongoing in both. Breathing rate is validated only during paced breathing at 6 breaths per minute. Readings were seated and still; results should not be extrapolated to motion or ambulatory use. The two investigators are excluded from every figure on this page.
License The Vitals SDK
Available to health, wellness and research products on iOS and Android. Tell us about your use case and we'll share integration details, pricing and the full validation dataset summary.
contact@neurofit.app
Interim results, updated as the validation study grows. The NEUROFIT Vitals SDK is not a medical device and does not diagnose, treat or prevent any condition.
Balance your nervous system, reduce stress, and feel your best with NEUROFIT.
Nervous System Fitness®
Our Products
NEUROFIT Mobile AppNEUROFIT Vitals SDKNEUROFIT For MilitaryNEUROFIT For Hospital StaffNEUROFIT For Teams/OrganizationsTools For Practitioners
Nervous System Dysregulation QuizcoachOS™ DashboardNEUROFIT Nervous System Book (PDF)Nervous System + Somatics GlossaryPersonalized Somatic Exercise ToolTop Glossary + Blog Articles
Research
Stress Outcomes WhitepaperBusiness Details
About UsApp EfficacyNEUROFIT App DemoWall Of Love 🤍Terms of ServicePrivacy PolicyXama Technologies, Inc. (dba NEUROFIT)
651 North Broad Street, Suite 206
Middletown, DE, USA 19709
SAM.gov Registration:
UEI: C6T8P1MJU3D5
CAGE: 19ML0
DUNS: 117864242
care@neurofit.apppress@neurofit.appcontact@neurofit.app© NEUROFIT 2026•All Rights Reserved
NEUROFIT®, Nervous System Fitness® and Coach Copilot® are registered trademarks of Xama Technologies, Inc. in the United States and other jurisdictions.
The NEUROFIT products, programs and services are intended solely to help users maintain a state of wellness and encourage a healthy lifestyle. NEUROFIT products, programs and services are neither designed nor intended for use in the diagnosis, cure, management, mitigation, prevention, or treatment of any disease or medical condition.
As always, we highly recommend that speak with your trusted healthcare professional before making any health-related decisions or using any health-related products or services.