What to know before your first night.

On a stable nightstand 1.5–3 feet (50–90 cm) from your head, microphone facing toward you, plugged in, with Do Not Disturb on. Don't put it under a pillow or blanket.

Keep the phone plugged in on a stable nightstand. Battery use varies with the phone, settings and recording duration, so a fixed percentage would not describe every overnight session.

Another person can affect a room-audio recording. Place the phone closer to the person being recorded and listen to the context when reviewing a candidate. SomniSense does not separate two sleepers or prove whose sound was captured.

Yes. Record with your bedside phone, then open the morning Night Report to review possible breathing pauses: estimated time and duration, marked waveform and surrounding sounds. Some nights contain no candidates, and the phone can miss events. More nights add context; they do not turn the result into a diagnosis.

PDF reports are currently available in English. Free has a 2-part report and Pro has a fuller 6-section report; physical page count varies with the available evidence. These are wellness reports, not clinical sleep reports.

No. Tagging is optional. You can skip the pre-sleep dialog and SomniSense will still monitor and produce a normal report. But the Lifestyle Lab™ Factor Impact analysis only works if you tag.

The published benchmarks report 91.67% sensitivity and 89.01% precision for 1-second snore-event detection (five-seed means), plus 88.49% accuracy and 88.06% F1 for the compressed breathing model on 200-second windows. These are cohort-level event/window results from 80 person-nights across 40 participants, not the accuracy of an individual night’s SRI or BRI and not a diagnosis. Use the accuracy and methodology link on this page for the full explanation.

BRI is a product-defined rate from bedside phone audio, not a clinical event count. Phone placement, room noise, estimated sleep time, model error, and the fact that audio does not measure airflow can all affect it. A difference from your expectation does not prove undercounting or overcounting.

No. SomniSense is a wellness and self-monitoring tool, not a medical device. BRI is a product-defined rate of candidate acoustic patterns per estimated sleep hour; it is not clinical AHI and clinical severity thresholds should not be applied to it. Diagnosis requires appropriate clinical testing and a qualified clinician's judgment.

The current technical minimums are iOS 16 or later and Android 9 or later. Device compatibility can also depend on the app stores and hardware available to your account.

Yes. Open a candidate and follow the playback pointer through the sound before the quiet interval and the sound that resumes. Room noise or another person may also be audible. Access and retention limits apply; deleted or expired recordings cannot be played. Sound alone does not confirm a medical event.

No, not by default. All AI analysis happens on your device. Audio only leaves your phone if you explicitly tap Export or Share.

Sleep audio is analyzed on your phone. Account and subscription information, aggregated sleep statistics, and technical context may be sent to provide account history, trends, reports and support. Raw audio is not uploaded by default. The privacy policy linked on this page gives the full inventory and your rights.

Yes. In Profile → Data Management, you can delete every recording, every report, and your account, anywhere, anytime.

Use the privacy policy link on this page to review the current data inventory, your rights and how to make a request.

No. We don't sell data or share it with advertisers.

Any trial depends on your purchase provider and eligibility. Before accepting, review the amount due today, duration, next charge date and cancellation deadline. Free remains available without a Pro purchase.

Check your chosen store or web checkout for current currency, tax, amount and renewal terms. Prices and offers can differ by provider and market. Start with Free if you want to review the first-night core evidence before deciding.

Use the provider shown on your receipt: Apple or Google subscription settings, or the web billing portal. Check its cancellation deadline and access end date, and keep the confirmation.

SomniSense cannot interpret an individual value medically, and BRI should not be mapped to AHI severity bands. If you have symptoms, concerns, or a persistent pattern that worries you, seek professional advice regardless of the number. Your multi-night record may provide useful context for that conversation.

The model has not been validated for users under 18. We don't recommend it for children.

SomniSense can add an independent bedside audio record that you review separately from your CPAP data. It cannot determine mask fit, residual AHI, treatment effectiveness, or whether therapy should change. Your CPAP data and clinician remain authoritative for treatment decisions.