Google Antigravity Browser Agent: Web Testing & Automation with Gemini 3
Google Antigravity Browser Agent: Web Testing & Automation with Gemini 3
Google Antigravity's Browser Agent is a game-changer for web automation. It combines Gemini 3's understanding with actual browser control, enabling intelligent testing, scraping, and interaction without writing complex scripts.
What is the Browser Agent?
The Browser Agent allows Gemini 3 to:
- Navigate websites - Browse like a human user
- Interact with elements - Click, type, scroll, hover
- Extract data - Scrape content intelligently
- Take screenshots - Capture visual evidence
- Run tests - Verify functionality automatically
Setting Up Browser Agent
Enable Browser Agent
In your Antigravity settings:
{
"antigravity.browserAgent.enabled": true,
"antigravity.browserAgent.headless": false,
"antigravity.browserAgent.defaultTimeout": 30000
}
Using with Playwright MCP
For advanced automation, connect Playwright MCP:
{
"mcpServers": {
"playwright": {
"command": "npx",
"args": ["@anthropic/mcp-server-playwright"],
"env": {
"PLAYWRIGHT_BROWSERS_PATH": "0"
}
}
}
}
Basic Browser Operations
Navigation
You: "Go to github.com and search for 'react native' repositories"
Browser Agent:
- Opens https://github.com
- Locates search input
- Types "react native"
- Submits search
- Reports results
Data Extraction
You: "Extract all product names and prices from this e-commerce page"
// Generated extraction script
const products = await page.$$eval('.product-card', (cards) => {
return cards.map((card) => ({
name: card.querySelector('.product-name')?.textContent?.trim(),
price: card.querySelector('.product-price')?.textContent?.trim(),
url: card.querySelector('a')?.href,
}));
});
console.log(`Found ${products.length} products`);
return products;