Google Antigravity is Google’s agent-first development platform. Instead of an assistant attached to an editor, it organizes work around agents that plan, execute and verify engineering tasks, with desktop, command-line and SDK surfaces.
What it’s for
- Delegating engineering tasks to agents that operate on the repository
- Working with subagents and scheduled flows for recurring tasks
- Creating and reviewing changes with a verification trail, such as tests and artifacts
- Extending the environment with your own extensions and integrations
- Carrying the same agent model into managed and enterprise scenarios
How it works
- The platform combines a desktop application, CLI, extensions, SDK and managed agents
- Agents receive tasks, plan, use tools and produce verifiable artifacts
- Worktrees and isolated environments let several agents work in parallel without colliding
- Scheduled tasks cover recurring routines, and the SDK lets you embed agents into other flows
Where it runs
- Desktop application
- Command line (Antigravity CLI)
- Extensions and SDK for integration
- Managed agents in the cloud, including an enterprise offering
Availability and licensing
- Proprietary Google product, subject to platform-specific terms of use
- Continuously evolving; offers individual plans and an enterprise version
- Takes the place of Firebase Studio as Google’s development offering
Strengths
- Agent-first approach with subagents and scheduled tasks
- Multiple surfaces (desktop, CLI and SDK) over the same agent model
- Integration with the Google ecosystem and with an enterprise offering
- Isolated worktree environments for parallel work
Things to note
- It is a distinct product from Gemini CLI and Jules; each covers a different flow
- Proprietary product subject to changes in availability and plans
- Adopting agents requires reviewing permissions and code governance