Google 推出了 Antigravity，一种新颖的 Agentic 开发平台，旨在提升开发者工作流程，使其更贴合任务需求。它将熟悉的 AI 驱动的编码环境（编辑器视图）与专用的“管理器界面”集成在一起，用于编排自主 Agent。这些 Agent 可以跨编辑器、终端和浏览器计划、执行和验证复杂的软件任务，从而减少频繁的上下文切换。实际应用包括委托多工具任务（如功能开发和测试），请求 UI 变更（Agent 提供可视化的 Artifacts 以供审查），以及派遣 Agent 进行后台维护或缺陷修复。一项关键创新是使用 Artifacts（屏幕截图、录像）来验证 Agent 的工作，相比原始日志，审查过程更直观且可信。Antigravity 旨在成为 Agent 驱动的开发中心，支持主要的 LLM，如 Gemini 3 Pro、Claude Sonnet 4.5 和 GPT-OSS，个人用户可在 MacOS、Windows 和 Linux 平台上体验公开预览版。
Development is lifting off. The tools of yesterday focused on helping you write code faster; the tools of tomorrow need to help you orchestrate it. Today, we’re introducing Google Antigravity, a new agentic development platform designed to help you operate at a higher, task-oriented level.
Antigravity isn't just an editor—it's a development platform that combines a familiar, AI-powered coding experience with a new agent-first interface. This allows you to deploy agents that autonomously plan, execute, and verify complex tasks across your editor, terminal, and browser.
A new way to work
We built Antigravity because we believe agents shouldn't just be chatbots in a sidebar; they should have their own dedicated space to work. The platform introduces two distinct ways to interact with your code:
- The Editor View: When you need to be hands-on, you get a state-of-the-art, AI-powered IDE equipped with tab completions and inline commands for the synchronous workflow you already know.
- The Manager Surface: This is where the shift happens. It’s a dedicated interface where you can spawn, orchestrate, and observe multiple agents working asynchronously across different workspaces.
Practical use cases for your workflow
Antigravity allows you to offload end-to-end tasks that previously required constant context switching. Here are three ways you can apply it to your daily development:
- Delegate complex, multi-tool software tasks to the agent: The agent can autonomously plan and execute the task across the editor, terminal, and browser, for instance, by writing code for a new feature, using the terminal to launch the application, and then using the browser to test and verify that the new component is functioning as expected, all without synchronous human intervention.
- Operate at a higher, task-oriented level by requesting UI changes: The agent will autonomously modify the codebase (UI iteration) and then communicate its progress and results via Artifacts, such as screenshots and walkthroughs, which provide the necessary and sufficient context for you to easily review and validate the agent's work and build trust.
- Utilize the agent-first Manager surface to dispatch agents for long-running maintenance tasks or bug fixes in the background: This capability enables you to delegate entire tasks, such as reproducing an issue, generating a test case, and implementing a fix, allowing you to focus on your primary work while observing the agent's progress asynchronously.
Verify with Artifacts, not logs
Delegating work to an agent requires trust, but scrolling through raw tool calls is tedious. Antigravity solves this by having agents generate Artifacts—tangible deliverables like task lists, implementation plans, screenshots, and browser recordings.
These Artifacts allow you to verify the agent's logic at a glance. If something looks off, you can leave feedback directly on the Artifact—similar to commenting on a doc—and the agent will incorporate your input without stopping its execution flow.
Built for choice and speed
We want Antigravity to be your home base for the era of agents. It treats learning as a core primitive, allowing agents to save useful context and code snippets to a knowledge base to improve future tasks.
Google Antigravity is available today in public preview, at no cost for individuals. This cross-platform solution is compatible with MacOS, Windows, and Linux, offering model optionality with generous rate limits on Gemini 3 Pro, and full support for Anthropic's Claude Sonnet 4.5 and OpenAI's GPT-OSS.
Download at antigravity.google/download and experience liftoff.
