Google launched another Coding Agent platform on Tuesday, this time focusing on developer teams collaborating to build agents that can execute complex tasks automatically, in a way that allows agents to be controlled remotely rather than actually being controlled independently.
The platform, called AntiGravity, is powered by Gemini 3 and is now available in public preview.”“Generous limits on the use of the Gemini 3 Pro,” Google writes in a blog post with the announcement.
AntiGravity is an agent coding platform that aims to build IDEs towards an agent-first future with browser control capabilities, seamless interaction patterns, and an agent-first product design.
Enterprises already burdened by the ever-increasing volume of code to review, thanks in large part to the rise of AI code generation, are demanding more from unobtrusive coding agents. They need unobtrusive coding agents to help developers review coding projects, review elements, and execute tasks autonomously.
For public preview, Integrity users can build agents using Gemini 3, Anthropic’s Sonnet 4.5 model, and Openai’s open source GPT-OSS. It will be compatible with developer environments running on major operating systems such as macOS, Linux and Windows.
“We want AntiGravity to be the home base for software development in the age of agents.” “Our vision is to ultimately enable anyone with an idea to experience liftoff and make that idea a reality.”
Google said it built Antigravity with four key principles.
Key principles of development
Enterprises today are either completely transparent about what’s going on under the hood, or they don’t show their work and isolate the code.
The AntiGravity team doesn’t think so either From these two extremes comes trust. "Antigravity provides context on agent work at a more natural task-level abstraction, with a necessary and sufficient set of samples and validation results, so that the user can have this confidence. A common emphasis is placed on thinking for the agent as a whole, not just the work, by affirming only the work for the agent." According to Google.
As for sovereignty, AntiGravity’s main interface, the Editor View, simulates an IDE experience, and standardizes what an agent might encounter as it accomplishes its tasks. The agent embeds this interface so that it can navigate to it.
However, Google plans to add “an agent-first manager layer” that flips that idea, meaning the interface is embedded in the agent.
The AntiGravity team created user feedback at “each level or pattern,” which would automatically feed into the agent’s execution. This will allow work to continue without humans needing to stop work to redirect the agent.
Iterate with the agent along with the human developer, "Self improvement" It becomes very important. Its agent can tap the knowledge base to learn from past work or contribute to new learning.
Google’s many coding agents
AntiGravity isn’t Google’s only coding platform. It’s also not the only coding agent with IDE integration or native capabilities. It joins a long line of Google platforms aimed at helping developers work more efficiently. The coding assistant Jules is now integrated into IDES, can be invoked via the CLI, and can also run asynchronously. The Gemini CLI works similarly. And there’s Gemini Code Assist, which first launched last year.
However, probably. Integrity will have to compete more with coding agent platforms such as Codex from OpenAI, Cloudcode from Anthropic, and Cursor.
Some people on X commented that antigravity looks like windsurfing, which would make sense. Google hired the Windsurf team — including CEO Varun Mohan — in July and Licensed the tech for $2.4 billion. Varun Mohan tweeted that this came from his team:
So far, early AntiGravity users have had mixed experiences, with many pointing to errors and slow generation.
Editors note: This story was updated on November 18, 2025 to include more information.