Breaking: Google is rolling out Google Antigravity, a new agent-first development platform that reimagines the IDE around autonomous AI agents powered primarily by Gemini 3 Pro, now in free public preview on Windows, macOS, and Linux for developers worldwide.
At its core, Antigravity looks like a familiar editor, terminal, and browser combo, but the real focus is long-running agents that plan tasks, edit code, install dependencies, run tests, and iterate while you supervise at a higher level. Google describes it as the home base for software development in the era of agents, turning high-level goals into multi-step workflows that can run well beyond a single prompt.
The product introduces two main views:
- Editor View: Keeps a traditional coding surface with an agent panel for inline suggestions and commands.
- Manager View: Lets you orchestrate multiple agents and workspaces in parallel, monitor progress, and resume task queues.
To address trust, Antigravity agents produce artifacts such as task lists, implementation plans, walkthroughs, screenshots, and browser recordings instead of only raw tool calls so developers can audit how results were produced.
Antigravity is positioned as a key pillar in Google’s Gemini 3 ecosystem, sitting alongside AI Studio, Vertex AI, and enterprise offerings as the place where day-to-day coding with agents actually happens. The platform is built on a Visual Studio Code foundation and is designed to support multiple frontier and open models in addition to Google’s own stack, signaling a push to win over professional teams with large, complex codebases. Early developer reactions describe it as an agent-focused reboot of the modern IDE that raises expectations for what autonomous coding tools should do out of the box.
