Antigravity Browser Extension: Complete Guide

Everything you need to know about installing, configuring, and using the Google Antigravity Chrome extension for agent-driven browser automation, UI testing, and web development.

Last updated: March 8, 2026 · 12 min read

What is the Antigravity Browser Extension?

The Google Antigravity browser extension is a Chrome extension that extends the capabilities of the Antigravity IDE beyond the code editor and terminal. It allows AI agents to directly interact with, test, and verify web applications running in Google Chrome, creating a seamless development workflow where agents can write code, launch it, and validate functionality in the browser — all autonomously.

The extension is optional but highly recommended for web development tasks. It enables automated UI testing, visual regression detection, interactive debugging, screenshot capture, session recording, and DOM analysis. Think of it as giving your AI agent eyes and hands in the browser.

Key benefit: The browser extension transforms Antigravity from a code-only AI assistant into a full-stack development partner that can see, interact with, and test your web application just like a human tester would.

How the Browser Subagent Works

When the browser extension is active, Antigravity spawns a browser subagent — a specialized AI model optimized for interacting with web pages. This subagent has access to a variety of tools for controlling your browser:

Interaction Tools

• Clicking buttons and links
• Scrolling pages
• Typing into input fields
• Selecting dropdown options
• Hovering over elements

Reading Tools

• DOM capture and analysis
• Screenshot capture
• Markdown page parsing
• Console log reading
• Network request inspection

Recording Tools

• Video recording of sessions
• Step-by-step screenshot sequences
• Action timeline capture
• Performance metrics logging

Verification Tools

• Visual regression comparison
• Element presence assertions
• Text content verification
• Layout and styling checks

When the agent is actively controlling the browser, you'll notice a glow around the browser window indicating the Antigravity Agent is in control. This visual indicator ensures you always know when an agent is interacting with Chrome.

Installation Guide (Step-by-Step)

Installing the Antigravity browser extension takes less than 5 minutes. Follow these steps:

Launch Antigravity IDE

Open Google Antigravity and start a new conversation in the Playground or Agent Manager. You need a personal Gmail account to use Antigravity (Workspace accounts are not supported yet).

Don't have Antigravity yet? Follow our getting started guide to download and install it first.

Trigger Browser Extension Setup

The first time an agent needs browser access, a Setup or Deny button will appear in the progress panel. Click "Setup" to proceed.

Alternatively, click the Chrome icon in the Antigravity toolbar at any time to manually trigger the extension setup flow.

Install from Chrome Web Store

You'll be redirected to the Chrome Web Store. Click "Add to Chrome" and approve the permissions dialog. The extension requires:

Active Tab — read and interact with the current tab
Screenshots — capture visible tab content
Native Messaging — communicate with Antigravity IDE

Verify Connection

Return to Antigravity IDE. A green indicator in the status bar confirms the extension is connected. If you see a yellow or red indicator, see the troubleshooting section below.

Test the Extension

Try a simple browser task: ask the agent "Open google.com and take a screenshot". You should see the agent open a browser window, navigate to the page, and produce a screenshot artifact.

Manual Installation & Recovery

If the automatic setup doesn't work, or if you've previously uninstalled the extension and need to reinstall it, follow these manual steps:

Fresh Manual Install

Open Chrome and navigate to the Chrome Web Store
Search for "Google Antigravity"
Click "Add to Chrome" on the official extension (publisher: Google LLC)
Restart Antigravity IDE to detect the newly installed extension

Reinstall After Uninstall

Go to chrome://extensions in Chrome
If the Antigravity extension is listed, click "Remove" to fully uninstall it
Restart Chrome completely (close all Chrome windows)
Reopen Antigravity IDE and click the Chrome icon in the toolbar
Follow the setup flow to install from the Chrome Web Store

Reset Extension Data

If the extension is installed but behaving unexpectedly, try resetting its data: go to chrome://extensions, find the Antigravity extension, click "Details", scroll down and click "Clear data", then restart Chrome.

Using the Extension with WSL

If you're running Antigravity IDE in WSL (Windows Subsystem for Linux), the browser extension works with Chrome installed on your Windows host machine. Here's what you need to know:

Requirements

• WSL 2 (WSL 1 is not supported for browser integration)
• Google Chrome installed on Windows (not inside WSL)
• Antigravity browser extension installed in the Windows Chrome
• Network connectivity between WSL and Windows host

Setup Steps for WSL

Install the Antigravity extension in your Windows Chrome browser
Launch Antigravity IDE from within your WSL terminal
Antigravity will automatically detect and communicate with the Windows Chrome instance via Native Messaging
When the agent opens a browser, it will open in your Windows Chrome

Tip: If Antigravity can't detect your Windows Chrome from WSL, ensure the BROWSER environment variable is set to your Windows Chrome path, e.g., export BROWSER="/mnt/c/Program Files/Google/Chrome/Application/chrome.exe"

Key Features & Capabilities

Automated UI Testing

Agents can navigate web pages, click buttons, fill forms, and verify expected behavior. Perfect for end-to-end testing and validation workflows. The agent clicks through your app like a real user and reports any issues.

Screenshot Capture & Visual Regression

Agents automatically capture screenshots at key points during testing. These become artifacts you can review to verify visual correctness. Screenshots can be compared across code changes to detect unintended visual regressions — catching CSS bugs and layout issues automatically.

Video Session Recording

Record entire browser sessions showing agent interactions. The browser subagent can capture video recordings of its actions, producing artifacts that show exactly what happened during testing. This is invaluable for debugging complex multi-step workflows, reproducing bugs, and sharing evidence with team members.

DOM Inspection & Content Parsing

Agents can read and analyze the full DOM structure, extract data, verify element presence and properties, read console logs, and even convert entire pages to markdown for analysis. This goes beyond what traditional testing tools offer, giving agents deep understanding of your page structure.

Real-Time Page Interaction

The Chrome extension integration allows agents to interact with web applications just like human testers — clicking buttons, filling forms, navigating pages, and verifying functionality. The browser becomes both a preview tool and a testing environment, catching UI/UX issues that traditional static analysis would miss.

Agent Manager Integration

The browser extension works seamlessly with Antigravity's Agent Manager, enabling powerful parallel workflows. Here's how they work together:

Parallel Browser Testing

Spawn multiple agents from the Agent Manager, each interacting with different browser tabs simultaneously. For example, one agent can test your login flow while another tests the checkout process — all running in parallel.

Build-Test-Verify Pipeline

Create a workflow where Agent 1 writes code, Agent 2 runs the dev server, and Agent 3 opens the browser to visually verify the changes. The Agent Manager orchestrates this pipeline automatically.

Cross-Browser Validation

Use multiple agents to test different viewport sizes and configurations simultaneously, ensuring responsive design works across mobile, tablet, and desktop breakpoints.

Learn more about the Agent Manager in our Agents guide and Features page.

Common Use Cases

UI Component Development

Agents build components, launch them in the browser, and verify they render correctly across different states and breakpoints. The agent can iterate on styling by seeing its own changes live.

Form Validation Testing

Test form submissions, validation logic, error messages, and success states automatically by having agents fill and submit forms with various valid and invalid inputs.

End-to-End Testing

Create comprehensive end-to-end test flows that simulate real user journeys through your application — from signup to checkout, and everything in between.

Bug Reproduction

Agents can reproduce reported bugs by following specific steps in the browser, capturing screenshots and video evidence for debugging.

Accessibility Auditing

Let agents navigate your application using keyboard-only interaction and inspect ARIA attributes to identify accessibility issues before your users encounter them.

Performance Monitoring

Agents can measure page load times, analyze console warnings, monitor network requests, and flag performance regressions introduced by code changes.

Permissions & Security

The extension requires the following Chrome permissions to function:

Active Tab:Read and modify content on the active tab. This allows the agent to interact with the web page you're developing.

Screenshots:Capture visible tab content for visual verification artifacts.

Native Messaging:Communicate with the Antigravity IDE process running locally on your machine.

Privacy & Security: The extension only operates when explicitly commanded by you or an agent you've authorized. It does not run in the background, does not collect browsing data, and all data stays local to your machine. The extension is published by Google LLC and is subject to the Chrome Web Store's security review process.

Troubleshooting

Extension not connecting to Antigravity

Ensure Antigravity IDE is running and you're signed in
Restart Chrome completely after installation (close all Chrome windows)
Check that the extension is enabled in chrome://extensions
Verify your firewall isn't blocking localhost connections on the port Antigravity uses
Try removing and reinstalling the extension (see Manual Installation)

Screenshots not capturing

Grant screenshot permission when prompted by Chrome
Ensure the target tab is visible and not minimized
Check available disk space (screenshots are stored temporarily)
Disable other Chrome extensions that might interfere with screen capture

Agent can't interact with page elements

Refresh the page after installing the extension
Some sites may block automation — check the console for CORS or CSP errors
Verify the extension has permissions for the target site
Check if the page uses Shadow DOM which may require different selectors
Ensure the page has fully loaded before the agent attempts interaction

Extension works on Windows but not in WSL

Ensure you're using WSL 2 (check with wsl --list --verbose)
Set the BROWSER environment variable to your Windows Chrome path
Check that networking between WSL and Windows host is working
Try restarting both WSL and Chrome

Video recording not working

Ensure sufficient disk space for temporary video files
Check that no other application is using the screen recording API
Try capturing a screenshot first to verify basic extension connectivity

Still having issues? Check our comprehensive troubleshooting page or visit the official documentation.

Frequently Asked Questions

Is the Antigravity browser extension free?

Yes, the Antigravity browser extension is completely free. It is part of the Google Antigravity IDE which is also free during the public preview period. No credit card or paid subscription is required.

Does the extension work on browsers other than Chrome?

Currently, the Antigravity browser extension only supports Google Chrome. Support for other Chromium-based browsers like Microsoft Edge and Brave may be added in the future. Firefox and Safari are not supported.

Can the extension access my personal browsing data?

No. The extension only operates when explicitly commanded by you or an agent you've authorized. It requires Active Tab, Screenshots, and Native Messaging permissions but does not collect, store, or transmit your browsing history or personal data. All data stays local to your machine.

How do I reinstall the extension if it stops working?

Go to chrome://extensions, find the Antigravity extension, click Remove, restart Chrome, reopen Antigravity IDE, and click the Chrome icon to trigger a fresh install. See the manual installation section for detailed steps.

Does the extension work with WSL?

Yes. When running Antigravity in WSL 2, the browser extension works with Chrome installed on your Windows host. The extension communicates via Native Messaging across the WSL boundary. See the WSL section for setup details.

What is the browser subagent?

The browser subagent is a specialized AI model within Antigravity optimized for interacting with web pages. It can click, scroll, type, read console logs, capture screenshots, record videos, parse DOM content, and convert page content to markdown. It runs when the main agent needs to perform browser-based tasks.

Can I use the extension with the Agent Manager?

Yes. Agents spawned from the Agent Manager have full access to the browser extension. You can run multiple agents that interact with different browser tabs simultaneously, enabling parallel testing and verification workflows. See the Agent Manager integration section above.