Coinciding with its announcement of Gemini 3 as we speak, Google introduced the launch of a brand new agentic improvement platform, Google Antigravity.
The corporate sees Antigravity as an evolution of the IDE into an agent-first future, with capabilities like browser management and asynchronous interplay patterns.
“With fashions like Gemini 3, we’ve got began hitting the purpose in agentic intelligence the place fashions are able to operating for longer durations of time with out intervention throughout a number of surfaces. Not but for days at a time with out intervention, however we’re getting nearer to a world the place we interface with brokers at greater abstractions over particular person prompts and gear calls. On this world, the product floor that permits communication between the agent and person ought to feel and look totally different – and Antigravity is our reply to this,” Google wrote in a weblog put up.
In response to Google, this new IDE is predicated on 4 key tenets of collaborative improvement: belief, autonomy, suggestions, and self-improvement.
To facilitate belief, Antigravity produces Artifacts as it really works, that are in codecs which are extra simply verified by builders than uncooked software calls, corresponding to job lists, implementation plans, walkthroughs, screenshots, and browser recordings. “Brokers in Antigravity use Artifacts to speak to the person that it understands what it’s doing and that it’s totally verifying its work,” the corporate wrote.
For autonomy, Antigravity’s default editor view features a full IDE expertise, with Tab completions, in-line Instructions, and brokers within the facet panel. There may be additionally an agent-first Supervisor floor, which can be utilized to create, orchestrate, and observe a number of brokers throughout a number of workspaces. These two views are deliberately not grouped right into a single view, however quite designed as separate areas with consideration for simply switching between them, Google defined.
The corporate additionally believes that one of many failings of brokers is the shortcoming to iterate with them. It mentioned that if an agent can full 80% of a job, however doesn’t present a means for builders to supply suggestions, then they typically need to spend extra time than the agent saved to complete the final 20%. Antigravity permits builders to offer asynchronous suggestions throughout any floor and Artifact in a lot of methods, corresponding to Google Doc-style feedback on textual content Artifacts or select-and-comment suggestions on screenshots. The brokers then take this suggestions and incorporate it again into their execution with no need to be stopped.
Equally, Google constructed self-learning into Antigravity, permitting it to retrieve from and contribute to improvement information bases. It could actually be taught from previous work, together with specific data like code snippets or derived structure and summary data just like the collection of steps taken to finish a job.
Google Antigravity is out there now in public preview totally free, with entry to Gemini 3, Anthropic’s Claude Sonnet 4.5, and OpenAI’s GPT-OSS.
