Overview
Google Antigravity: Multi-agent orchestration
Google Antigravity is Google's agent-first development platform, built to move beyond simple code completion into full task delegation. Rather than acting as a chatbot in a sidebar, Antigravity gives AI agents their own dedicated workspace where they can autonomously plan, write, run, and verify code across the editor, terminal, and browser at once.
With Antigravity 2.0 (launched at Google I/O 2026), the platform expanded into four connected surfaces: a desktop app for multi-agent orchestration, a CLI for terminal-native workflows, an SDK/API for building custom agents, and the original editor-centric IDE view. A "Manager Agent" breaks a task into subtasks and runs specialized agents — for architecture, implementation, testing, and CI — in parallel rather than sequentially.
To build trust in autonomous work, Antigravity produces "Artifacts": task lists, implementation plans, screenshots, and browser recordings that let developers review an agent's logic without digging through raw tool calls. It supports model optionality, including Gemini 3.5 Flash, Claude, and other models, and runs on macOS, Windows, and Linux.