Google Antigravity

Run multiple coding agents in parallel and review their plans, artifacts, and diffs

AIコーディングIDECoding AgentAgent-First IDEParallel AgentsTrial

Google Antigravity is an agent-first development platform with a desktop app, IDE, CLI, and editor extensions. It runs multiple coding agents in parallel, produces reviewable plans and artifacts, and supports frontier Gemini, Claude, and GPT models on a free individual tier with paid Pro and Ultra plans.

公式サイトへ

更新日

2026-10-21

対応プラットフォーム

Windows / macOS / Linux / Web

お試し

Freemium: the Individual plan is $0 per month with access to a variety of agent models, unlimited tab completions, unlimited command requests, and basic weekly rate limits. Google AI Pro and Ultra add more rate limits and a flexible AI credit pool; Business is available through Google Cloud, with Standard/Plus plans from $30 per seat per month or pay-as-you-go consumption pricing. Confirm current limits and model access on the official pricing page.

* 料金は参考情報です。最新の料金は公式サイトをご確認ください

antigravity.google

Google Antigravityの主な機能

Parallel Agents
Manage multiple local agents in parallel across Projects, workspaces, and background subagents.
Artifacts & Verification
Inspect implementation plans, walkthroughs, screenshots, and verification tests before accepting changes.
Multi-surface Workflow
Work in the Antigravity App, Antigravity IDE, CLI, or extensions for VS Code, JetBrains, Zed, and Xcode.
Free Individual Tier
Start at $0 per month with agent model access, unlimited tab completions, and unlimited command requests.

Google Antigravityのおすすめ用途

Enterprise developers in large codebases

Run multiple agents across a large repository while reviewing plans, diffs, and verification artifacts.

Full-stack developers

Move from prototype to production with designed artifacts and comprehensive verification tests.

Frontend developers

Use browser-in-the-loop agents to automate repetitive UX development and verification.

長所

Free individual plan includes agent models, unlimited tab completions, and unlimited command requests
Runs multiple agents in parallel with projects, workspaces, and subagents
Produces reviewable plans, artifacts, screenshots, and verification tests
Available as a desktop app, IDE, CLI, and extensions for major editors

注意点

Higher usage depends on Google AI Pro or Ultra rate limits and AI credit pools
Business plans require Google Cloud / Gemini Enterprise onboarding
Model availability and quotas vary by plan

Google Antigravity のクイックスタートと実践のコツ

クイックスタート

1
Give each parallel agent a different slice of the task, such as one module or one bug, instead of running the same prompt twice and comparing.
2
Treat the plan and verification output as the review surface: approve the plan before the agent writes, then read the artifacts rather than re-reading every diff line.
3
Keep the IDE for interactive work and the CLI for repeatable runs, so the same review habit covers both.

実践のコツ

Parallel agents only pay off if each one has an isolated scope and a clear finish line. Two agents editing overlapping files produce conflicts you then have to resolve by hand, which costs more than running them in sequence. Because the individual tier is free with model availability tied to the plan, validate your workflow there before paying, and recheck which models your plan exposes when output quality suddenly changes.