“With TrulySecure Sound ID, we’re making it possible for companies to create smart always-listening products that can alert us of things happening within our homes that we may not hear, and safeguard our home and family from potentially dangerous situations.” Todd Mozer, CEO, Sensory

A new kind of home security solution is the product of Sensory‘s latest adaptation of its industry-leading wake word technology: A system designed to listen for signs of trouble such as breaking glass or a crying baby.

Part of Sensory’s TrulySecure Speaker Verification suite of AI technologies for smart speakers and other Internet of Things devices, the solution is called Sound ID. It uses the same fundamental principle underlying Sensory’s wake word solutions, which are always listening for the keywords used to fully activate a given device, but instead uses this approach to listen for environmental sounds that may be of interest to a user. That way, a doorbell ringing or a smoke alarm going off can automatically send an alert to a user, prompting them to take action.

What’s more, the solution is designed to operate entirely on a given device, and not in the cloud. That means the audio data collected won’t compromise the privacy of the user; it isn’t sent to an external server, but rather is processed within the device itself.

In a statement announcing Sound ID, Sensory CEO Todd Mozer explained that with Sound ID his firm is enabling companies that integrate it “to create smart always-listening products that can alert us of things happening within our homes that we may not hear, and safeguard our home and family from potentially dangerous situations.”