HRV API: How to Get Heart Rate Variability Data

Heart rate variability (HRV) is a measured metric derived from the tiny timing differences between consecutive heartbeats (the RR or interbeat intervals). You get it two ways: from an on-device store — Apple HealthKit on iOS, Android Health Connect on Android — or from a cloud wearable API (Oura, WHOOP, Garmin, Fitbit) after the user's device syncs. A key catch is that providers report different statistics: Apple stores SDNN, while Health Connect and most others report RMSSD, and the two are not directly comparable. Best pick: for recovery/readiness scoring out of the box, Oura or WHOOP; for a raw RMSSD trend on Android, Health Connect; for real-time biofeedback, a chest strap's RR intervals. (For what HRV actually represents, see what is HRV.)

Where you can get HRV

HRV needs a wearable — a phone alone cannot measure it. All provider-specific fields, statistics, and device coverage change often, so treat this table as a starting point and confirm against each vendor's live data dictionary (as of 2026, verify).

Strava does not expose HRV.

Measured or estimated?

HRV is measured — the underlying beat-to-beat intervals are real timing data. What must be hedged is the interpretation. RMSSD reflects parasympathetic (vagal) influence on heart rate; it is a correlate, not a direct readout of "vagal tone," "stress," or "recovery." Do not present HRV as something that measures your nervous system or diagnoses anything. Readiness and recovery scores built on HRV (Oura, WHOOP, and others) are proprietary vendor models, so describe them as estimates, not clinical measures. Consumer HRV is a general wellness signal, not a medical diagnostic.

Two accuracy and comparability gotchas to design around:

• RMSSD vs SDNN are different statistics and are not directly comparable. Apple gives you SDNN; Health Connect, Fitbit, and most others give RMSSD. Label which one you are storing and never mix them in one trend line.
• Measurement window matters. Most consumer platforms compute HRV over a nocturnal/overnight window for comparability (Oura, WHOOP, Garmin, Fitbit-during-sleep); some apps instead use a guided morning spot reading. Night and morning values are not interchangeable — surface which window a reading came from. And because absolute HRV varies enormously between people, it is meant for within-person trends, not cross-person comparison. Verify which statistic and which time window each provider returns before you normalize.

Which source should you pick?

• Recovery/readiness scoring out of the box: Oura or WHOOP — nightly HRV is already baked into their readiness/recovery models, so you consume a score rather than build one.
• Raw RMSSD trend across Android devices: Health Connect HeartRateVariabilityRmssdRecord, read on-device for privacy and no post-sync latency.
• iOS-native app: HealthKit HKQuantityTypeIdentifierHeartRateVariabilitySDNN — but remember it is SDNN, not RMSSD.
• Real-time HRV (biofeedback): a chest strap's RR intervals over BLE/SDK; cloud wearable data is latent and lands only after the device syncs.
• Many brands at once: an aggregator (Terra, Junction, Rook) hands you one normalized HRV schema instead of N integrations. If you are weighing rings against straps, see Oura vs WHOOP.

Before you ship

Field names, statistics (RMSSD vs SDNN), scopes, and device coverage are volatile — re-verify each against the vendor's current data dictionary as of 2026. Label the statistic and the measurement window on every stored value, and frame HRV as a within-person wellness signal that reflects vagal activity, not a diagnostic or a direct measure of stress or recovery.