Skip to main content

Somnus Technologies

Nightsong provides a phone-based sleep monitoring solution that captures and analyzes nightly audio without requiring a wearable device. The app processes all audio on-device, scores sleep quality against a validated library of expert-labeled recordings, and discards raw audio so only de-identified metrics leave the phone. It is currently available in beta.

Traction 1/10
Milwaukee, United States · HQ
Founded 202510+ followers
  • Digital Health
  • Healthcare Technology
  • Wellness & Fitness Technology
  • Software Only
Updated 9 days ago

Funding

Funding rounds are not available yet.

Founders

Founder details are not available yet.

Product

Problem

Traditional sleep monitoring requires either wearable devices that can be uncomfortable or intrusive, or clinical sleep studies that are expensive and inconvenient. Many people lack access to actionable feedback about their sleep quality, leaving a third of their lives unexamined and unoptimized.

Solution

Nightsong provides a phone-based sleep monitoring solution that captures nightly audio from the phone placed by the bed, requiring no wearable device. The app uses on-device processing to score sleep quality against a validated library of 287 expert-labeled overnight recordings, with 60 held-out recordings used to verify every engine change. Raw audio is discarded on the device, ensuring only de-identified numbers leave the phone. When the user consents, Nightsong delivers actionable recommendations for improving sleep, all while maintaining strict privacy with no ads or data sales.

Target Audience

Primary users are health-conscious consumers who want to understand and improve their sleep quality without wearing a device or visiting a sleep clinic, as well as wellness-focused individuals seeking privacy-preserving health monitoring.

Features

  • On-device audio processing with no upload path for raw audio; audio is listened to, scored, and discarded on the phone
  • Sleep scoring validated against 60 held-out sleep-lab recordings from an outside library of 287 expert-labeled overnight recordings
  • Privacy-first architecture with no ad or analytics SDKs, no data sales, and exports that leave audio out
  • Share cards contain only one night, one Index, and one card with no underlying data traveling with it
  • De-identified numeric data transmitted to secure company-owned storage with no name or email attached
  • Consent-based actionable recommendations delivered only when the user opts in
This profile is AI-generated and may contain inaccuracies.