$ buildwith info google-antigravity

Google Antigravity

#6 in AI Code Editors · Freemium

sources differ by section

Google Antigravity is Google's agentic development platform. It ships as several surfaces that share one agent harness: the Antigravity IDE (a full editor with tab completion, natural-language commands and an Agent Manager), the Antigravity App (a command center for running multiple local agents in parallel across projects), a terminal-first CLI, extensions for other IDEs, and a Python SDK for building custom agents on the same harness. It runs on macOS, Windows, Linux and Googlebook, and there is also a web app.

for Google pitches it at everyone from enterprise developers in large codebases to hobbyists vibe coding in their spare time. Its use-case pages cover frontend and full-stack work, legacy migrations, automated testing and QA, incident triage and onboarding, plus non-engineering roles such as product and design.

use this if you want Google's agent-first IDE with a separate app for running many agents in parallel, browser-in-the-loop checks, and Gemini models included at no charge to start.

open antigravity.google ↗

Features

$ features --list

  • ›Agent that works across the editor, terminal and browser, so it can check its own UI work in a real browser
  • ›Artifacts: the agent reports progress as reviewable deliverables, and you can comment on any artifact to steer it
  • ›Agent Manager in the IDE for longer, multi-surface tasks
  • ›Antigravity App: run agents in parallel, group conversations into Projects that span folders, with scoped permissions
  • ›Dynamic subagents and cron-style Scheduled Tasks that run in the background
  • ›Skills, MCP servers and JSON hooks, set globally or per workspace, plus a plugin marketplace
  • ›Live voice transcription for spoken prompts
  • ›Python SDK for prototyping custom agents and running evaluations on the Antigravity harness

Pricing

$ pricing --tiers

Individuals

$0/month

App, CLI, IDE extensions and Python SDK, with parallel background agents, subagents, browser use and scheduled tasks. Baseline agent quota refreshed weekly. Gemini 3.8 Flash included; the Claude Sonnet 4.6, Claude Opus 4.6 and GPT-OSS-120b models on this tier are marked for removal on November 30, 2026.

Google AI Pro

Google AI Pro plan

Quota refreshed every 5 hours with higher weekly limits, /boost deep reasoning, multi-agent teamwork, Remote Control and optional AI credit overages. Adds Gemini 4 Neon, Argon and Hydrogen, Orchestrator (Neon and Argon), and Claude Sonnet 5.5 and Opus 5.5 on paid (non-trial) plans. Requires a personal Google account; the price is set by the Google AI plan, not the Antigravity pricing page.

Google AI Ultra

Google AI Ultra plan

Highest quotas and weekly limits, adds Gemini 4 Carbon and Orchestrator (Carbon), with full third-party partner model access and AI credit overages.

Business via Google Cloud

From $30/seat/month, or pay as you go

API access with consumption pricing. Eligible Gemini Enterprise customers get Standard/Plus plans from $30 per seat per month, or a pay-as-you-go plan with a $0 seat fee. Spend caps and central admin controls; partner models need admin opt-in and bill by tokens.

Why this rank

$ diff --rank 6

Ranked #6 among AI code editors. We re-checked it on 2026-11-20 against the same six-part rubric as the other editors and it stays below Kiro. Builder fit (25%), the heaviest weight, is about staying hands-on with the code: Antigravity is now agent-first, with its parallel-agent App leading and the IDE as one surface, while Kiro keeps the most structured in-editor workflow (specs, checked steps, property-based tests). Adoption (20%) and community sentiment (10%) have no comparable measure for closed editors yet, and we have hands-on time (15%) with neither, so none of those can lift it. Momentum (15%) is fast for both. Pricing / value (15%) is where Antigravity leads: its free tier includes every agent feature, limited by a weekly quota, against Kiro's 50 credits a month. That is one lighter dimension, and being free is not enough to move it, while its paid tiers have no price on the pricing page and its model lineup changes week to week. Cursor, Devin Desktop and Copilot rank above both on how widely they are already used, and Zed stays the fast editor with your own keys. These ranks are editorial; the scores are unmeasured placeholders and are not shown.

Strengths

$ strengths

  • ›Free tier includes the full agent feature set (parallel agents, subagents, browser use, scheduled tasks), limited by quota rather than features.
  • ›Browser-in-the-loop agents and reviewable artifacts make it easier to check UI work than reading a diff alone.
  • ›One harness across IDE, desktop app, CLI, other-IDE extensions and an SDK, so the same agent setup follows you between surfaces.
  • ›Fast model cadence: Google's blog shows new Gemini 4 models and features landing in Antigravity roughly weekly through October and November 2026.

Limitations

$ limits

  • ›Quotas, not prices, define the paid tiers: the pricing page says rate limits and credit pools vary by plan and task complexity, and the Pro and Ultra prices are not listed there.
  • ›Google AI Pro and Ultra require a personal Google account; Workspace work accounts need pay-as-you-go through Google Cloud or Gemini Enterprise.
  • ›Model lineup on each tier changes often; the Claude 4.6 and GPT-OSS-120b models are scheduled to leave the plans on November 30, 2026.
  • ›The IDE and app are proprietary. Only the Python SDK is open source (Apache-2.0, google-antigravity/antigravity-sdk-python, 4,124 GitHub stars on 2026-11-20); the antigravity-cli repo has docs and a changelog but no licence.

Sources

$ sources --verified

$ buildwith alternatives google-antigravity

  • #1 CursorFreemium

    use this if you want a VS Code-style editor with AI deeply woven into every edit.

  • #2 Devin Desktop (formerly Windsurf)Freemium

    use this if you want the Windsurf IDE experience plus a command center for managing coding agents.

  • #3 GitHub CopilotFreemium

    use this if you want AI help inside the editor you already use, tied into GitHub.

  • #4 ZedFreemium

    use this if speed matters most and you want a fast native editor with AI built in.

  • #5 KiroFreemium

    use this if you want an AWS-run agentic IDE that turns prompts into specs, checks work with property-based tests, and can keep agents running in the cloud or from Slack.

sponsor this spot →