How it works
Four small things.
One quiet system.
Rezpio runs continuously in the background of your day. Here's what it actually does, and, just as importantly, what it doesn't.
Wear the Clip.
Clip Rezpio to your collar or lapel in the morning. It's smaller than a lozenge and quiet enough to forget. Seven days of battery, one charge on Sunday.
It listens, passively.
No breathing tests. No prompts. Rezpio captures the small acoustic signals in the way you already talk, laugh, and breathe, the ones a stethoscope can't hear over the length of a day.
Edge-AI finds the pattern, on the device.
The signal is processed by a dedicated on-device model trained on the world's largest voice-biomarker dataset. Only the derived patterns ever leave the Clip. Audio does not.
See your trends. Share when you're ready.
Your patterns build week over week in the Rezpio app. When something shifts, you'll see it early, and you decide when your care team sees it, if at all.
Why it's clinical
Three sensors. One respiratory signature.
Voice alone is powerful. Voice fused with a 3-axis accelerometer and a calibrated ambient temperature sensor is clinical. Every cough is reconstructed as a 3D biomechanical event, grounded in the environment it happened in.
3-axis accelerometer is currently highlighted in the signature above. All sensor streams are processed on-device, only derived patterns leave the Clip, never audio.
Your privacy
Your voice stays yours.
Always.
Rezpio is the only continuous respiratory monitor of its kind engineered so that recordings of your voice never leave the device. Not to our servers. Not to your care team. Not to anyone.
Audio in. Patterns out. Nothing else.
Processed on the Clip itself.
A dedicated neural chip inside Rezpio turns acoustic signal into biomarker features on the device, in milliseconds, then discarded.
Only derived patterns sync.
What reaches your phone is a compact numeric summary: respiratory rate variability, cough events, voice-quality indices. Never a recording.
You control who sees what.
Nothing is shared with a clinician, family member, or health system until you explicitly grant access, per-metric, revocable at any time.
- Records a 30-second audio clip of you speaking.
- Uploads the raw recording to a remote server.
- Processed, stored, and analyzed off-device.
- Retained under a data-retention policy you don't set.
- Listens continuously to your natural speech, all day.
- Extracts biomarker features on the Clip, in real time.
- Audio buffer is discarded within milliseconds.
- Only numeric patterns ever sync, you own them.
Rezpio is HIPAA-aligned and independently audited. Learn more in our Privacy Architecture whitepaper.