Sustained voice detection
Designed to alert after 30 seconds of sustained voice activity during configured quiet hours.
Quiet-hour monitoring with privacy in mind.
DataChill Quiet is a monitoring system under development for shared living environments. It is designed to detect sound events, not identify people or determine who caused a disturbance.
Pre-launch · Product under development
Planned capabilities
Loud impact detection
Designed to alert promptly when a sudden loud sound exceeds configured conditions.
Local processing
Production sound analysis is specified to occur on the monitoring device.
Privacy-first design
No speech transcription, speaker identification or intentional storage of raw audio.
Telegram notifications
Qualifying detection events are intended for a configured Telegram destination.
Configurable quiet hours
Quiet-hour schedules, thresholds, durations and silence-reset settings are configurable.
Installation by DataChill
Installation and device configuration are planned as part of the service.
Structured event history
The approved specification calls for up to 90 days of event metadata retention.
How detection is specified
The planned categories are QUIET, VOICE, LOUD IMPACT and SHORT SOUND. Brief ordinary sounds should be ignored. Voice alerts follow 30 seconds of sustained activity; qualifying loud impacts trigger an immediate alert. Thresholds, durations and silence-reset periods can be adjusted during calibration.
Detection accuracy depends on room conditions, device placement and calibration. Perfect classification or detection cannot be guaranteed. Prototype detection work is underway; the finished production device and commercial service have not been verified as available.
Future plans
WhatsApp alerts and a mobile app are being considered for a later phase. They are not part of the current Telegram-only specification, and no release date or availability is promised.
What happens to sound
The planned production design processes microphone input locally and does not record, save or transcribe raw audio. It produces structured detection events; Telegram delivery involves an external service. The intended event retention is 90 days. See the privacy policy for the distinction between audio, event metadata and website enquiries.
Ask about availability