Google adds cloud-based agents and voice controls to Antigravity
Remote Agents are rolling out to paid Google AI Pro and Ultra subscribers, allowing development tasks to continue after a laptop is closed.
Google is updating its Antigravity development platform with cloud-based Remote Agents, spoken interaction powered by Gemini Live and a redesigned interface. Version 2.29 is rolling out for Mac, Windows and Linux.
For paid Google AI Pro and Ultra subscribers, Remote Agents support projects created from a GitHub repository or Google Drive folder. Google says tasks run in separate cloud virtual machines and continue after users close their laptops. Users can manage the agents by phone through Antigravity’s existing remote website.
The new Voice Mode allows users to speak requests to an agent. It offers multiple voice options and adjustable speaking speeds ranging from 0.75x to 2x.
The conversation screen now places the prompt field at the bottom and a greeting above it. Google Sans becomes the default font, alongside the Onyx theme. For single-repository projects, branch and Git worktree choices share a menu, while agent selection moves into the model menu.
Other changes cover automation management and browsing. Automations are grouped by whether they are scheduled, always active or paused, with previous and upcoming run times displayed. The built-in browser gains download progress tracking, browsing-history suggestions and a selectable default download folder. Users can also customize keyboard shortcuts. Google says the update reduces cold-start launch times and smooths sidebar and side-panel animations.
