Choose your HRV source. Camera is convenient for rough trends. A Bluetooth HR sensor such as WHOOP or a Polar H10/chest strap is better for R-R interval biofeedback.
Setup: WHOOP users can enable Heart Rate Broadcast in the WHOOP app and keep the sensor snug and still. Polar H10 or chest-strap users should wet the electrodes thoroughly, snap the pod on, and wear it snug below the chest muscles. Make sure the sensor is not already locked by another app. Then click Connect - Chrome will show a Bluetooth picker. Web Bluetooth works in Chrome, Edge, Brave on desktop; not Safari/Firefox.
HF broadly tracks respiratory vagal modulation and often rises with slow breathing. LF is mixed sympathetic + parasympathetic and can dominate during paced breathing at 0.1 Hz (6 bpm). LF/HF is a contested proxy for sympathovagal balance — interpret cautiously. Rolling 60-second window per Task Force 1996 and ultra-short HRV validation literature — HF is reliable at this duration, LF stabilizes around 30s.
Calibration note. Absolute ms² values use an empirical scaling constant tuned to typical literature ranges. They're reliable for tracking change across your own sessions on the same device, but not directly comparable to clinical norms or other apps — each HRV tool uses its own PSD method and normalization. LF/HF ratio is not affected by that shared scaling constant, but still depends on windowing and preprocessing choices.
Capture a 30-second resting HRV reading before (and optionally after) your breathing session. Lets you see exactly how the practice moves your nervous system, separate from where you started.
Sit still, face the camera, good even lighting. HR settles in ~10s; HRV needs ~30–60s.
Try each rate as coherent breathing, then capture the trial to compare average rMSSD and coherence.
Choose a rate, press begin, breathe for 60–90 seconds, then capture.
Your past breathing sessions, stored locally on this device.
Local only. Session history stays in this browser's localStorage. CSV and JSON exports are generated on this device.
No sessions yet. Complete a 2+ minute breathing session and it'll appear here.
Optional cloud sync
You can keep using the app without an account. Signing in only adds synced history for this app.
5 minutes · coherent breathing · 30 cycles
ms² values use an empirical scaling constant — reliable for tracking your own change over time, not directly comparable to clinical norms. LF/HF ratio is not affected by the shared scaling constant, but still depends on windowing and preprocessing choices.
Session metrics CSV contains all the aggregate values shown above. R–R interval CSVs come from the connected Bluetooth HR sensor. ECG CSVs are available only when a Polar H10 exposes raw ECG data, ready for analysis in Python, R, or any tool that reads CSV.
Adjust the length of each phase to match your own rhythm.
Timed programs auto-stop at the chosen duration with a gentle ending chime. Open lets you stop manually whenever you're ready.
Flicker safety. This feature is off by default and is disabled automatically when your system requests reduced motion. Do not use rhythmic light stimulation if you have photosensitive epilepsy, a seizure history, migraine sensitivity, or if it makes you feel unwell.
Subtle sits below the comfort threshold — gentle background pulse, fine for long sessions. Optimal targets ~40% modulation depth on exhale, the saturation point where SSVEP neural drive peaks (Norcia et al., 2015). Past this, the brain can't respond more strongly even if the screen flashes brighter. Maximal keeps the same neural peak but drops inhale brightness to ~8%, giving a much larger perceptual contrast — feels more dramatic, same effective entrainment.
Alpha · 10 Hz drives calm-aware mindfulness, the eyes-open meditation state. Best-validated frequency in the SSVEP literature (Notbohm et al. 2016; Gulbinaite et al. 2017).
When on, the audio binaural beat tracks the visual flicker — strongest entrainment effects in the literature come from synchronized audio+visual stimulation at the same frequency (Jirakittayakorn & Wongsawat 2017; Cherry et al. 2025). Off lets you mix bands (e.g. alpha visual + theta audio).
Long wavelengths (red/amber) activate the melanopsin photoreceptors in the retina much less than blue light does. Lower melanopsin signaling → lower sympathetic arousal → easier parasympathetic shift. This is the same principle behind dim red lights for evening wind-down (Pilorz et al. 2016). Amber is the recommended default; Sunset is the strongest parasympathetic-favoring choice but reads almost monochrome.
Encourages a wide, soft gaze rather than focused stare. Stress narrows attention onto a focal point (Easterbrook 1959 and the half-century of follow-up work); the reverse — that deliberately widening visual attention downregulates arousal — is a popular extrapolation but rests more on the underlying neuroscience of the locus-coeruleus arousal system than on direct intervention trials. Treat as plausible-and-low-cost rather than RCT-proven.