Google Antigravity
Overview
Section titled “Overview”| Company | |
| Type | IDE (VS Code fork) |
| Pricing | Free tier (Gemini 3 Pro) / Google AI Pro & Ultra for more |
| Protocol | MCP, Skills |
| Config Formats | AGENTS.md, GEMINI.md |
| Website | antigravity.google |
What It Does
Section titled “What It Does”Google Antigravity is an agent-first IDE that combines a traditional coding experience with autonomous AI agents. It features two primary views: an Editor View (standard IDE with AI sidebar) and a Manager View (control center for orchestrating multiple agents working in parallel across workspaces).
Key Strengths
Section titled “Key Strengths”- Agent-first architecture: Spawn and manage multiple autonomous agents working in parallel
- Multi-model: Supports Gemini 3 Pro, Claude Sonnet 4.5, and GPT-OSS models
- Manager View: Orchestrate asynchronous agent workflows across workspaces
- MCP support: Connect external tools and data sources including Google Cloud services
- Artifacts: Agents produce verifiable deliverables like screenshots and browser recordings
- Browser integration: Agents can autonomously interact with browser and terminal
- Free tier: Generous free usage with Gemini 3 Pro during public preview
Pricing
Section titled “Pricing”| Plan | Price | Rate Limits |
|---|---|---|
| Free | $0 | Weekly quota, Gemini 3 Pro, unlimited code completions |
| Google AI Pro | Subscription | Higher rate limits, 5-hour quota refresh |
| Google AI Ultra | Subscription | Highest rate limits, 5-hour quota refresh |
All tiers get full feature access including Gemini 3 Pro, unlimited tab completions, Agent Manager, and Browser integration. Paid tiers differ only in rate limit generosity and refresh frequency.
Best For
Section titled “Best For”- Developers wanting an AI-native IDE with multi-agent orchestration
- Google Cloud projects and ecosystem users
- Teams managing complex, parallel coding workflows
- Developers who want multi-model flexibility in a single IDE