Google Home routines now support custom visual triggers

Google Home is getting some new capabilities courtesy of Gemini. Starting today, users signed up for early access to Gemini on Google Home are able to create automations that’ll start when compatible cameras detect more or less any type of activity that you can write into a prompt.

Google announced the new features today in a Nest Help forum post. The post gives a handful of examples of the new types of “visual insights” that Nest Cams and other Gemini Built-in cameras can now use as starting points for custom routines: Google says nearly any event can be a trigger, including things like Amazon deliveries, car doors left open, and raccoons spotted in certain areas.

Visual events can be used to start any Google Home routine, so you could have the app send you a notification when a camera sees you’ve left your car door open, or have floodlights flip on when a raccoon is spotted poking around your garbage bins. It seems like any visual condition that Gemini might be able to recognize through a camera feed can be used as a trigger.

In addition to the new visual capabilities, Google says the Gemini experience in Google Home is getting a few lower-key improvements, too. As of the latest update, Gemini should be more able to understand plain language commands like “make the living room a little warmer.” You’re also able to string multiple actions together in a single request more reliably, and Google says basic tasks like setting timers and alarms should feel a little faster.

Gemini aside, the Home app’s widgets should now be more responsive, Google says. You can sign up for Gemini in Google Home by opening the Google Home app, tapping your profile picture, then Home settings -> Early access.