browse

Browser automation with headless Chrome/Firefox for testing, scraping, screenshots, and interactive web operations with full JavaScript rendering.

Overview

The browse primitive provides powerful browser automation capabilities including page navigation, form interaction, screenshot capture, and PDF generation with Playwright/Puppeteer integration.

Quick Example

import { browse } from 'sdk.do'

// Navigate and screenshot
const screenshot = await browse.screenshot({
  url: 'https://example.com',
  selector: '#main-content',
  format: 'png',
})

// Interact with page
await browse.interact({
  url: 'https://example.com/login',
  actions: [
    { type: 'fill', selector: '#email', value: '[email protected]' },
    { type: 'fill', selector: '#password', value: 'secret' },
    { type: 'click', selector: '#login-button' },
    { type: 'waitFor', selector: '.dashboard' },
  ],
})

// Generate PDF
const pdf = await browse.pdf({
  url: 'https://example.com/invoice/123',
  options: {
    format: 'A4',
    printBackground: true,
  },
})

Core Capabilities

• Headless Browser - Chrome and Firefox automation
• Page Interaction - Click, fill, scroll, navigate
• Screenshot Capture - Full page or element screenshots
• PDF Generation - Convert web pages to PDF
• JavaScript Execution - Run custom scripts in browser

Access Methods

SDK

TypeScript/JavaScript library for browser automation

await browse.screenshot({ url: 'https://example.com', format: 'png' })

→ SDK Documentation

CLI

Command-line tool for browser operations

do browse screenshot https://example.com --output screenshot.png

→ CLI Documentation

API

REST/RPC endpoints for browser automation

curl -X POST https://api.do/v1/browse/screenshot -d '{"url":"https://example.com"}'

→ API Documentation

MCP

Model Context Protocol for AI-driven browsing

Take a screenshot of https://example.com and save it

→ MCP Documentation

Related Primitives

• scraper - Web scraping
• tests - E2E testing
• fetch - HTTP requests