Browser Scraper

@rafaelgdn/browser-scraper is a CDP-first browser scraping package for Node.js, preserving the same public API and the same design direction used in the original GDNox prototype.

Status

This package has the core ported:

• Direct Chrome launch with raw CDP
• Tab/page control
• Unified DOM search with shadow root piercing
• OOP iframe support
• Element interactions
• Network monitoring and interception
• Profile seeding

Install

pnpm install

If pnpm blocks native build scripts, run:

pnpm rebuild better-sqlite3 esbuild

Usage

import { Browser } from "@rafaelgdn/browser-scraper";

const browser = new Browser();

try {
  const tab = await browser.new_tab();
  await tab.goto("https://example.com");

  const heading = await tab.find("h1");
  console.log(await heading?.text());
} finally {
  await browser.close();
}

API

Browser

new Browser({
  chrome_path: null,
  headless: false,
  user_data_dir: null,
  proxy: null,
  extra_args: [],
  auto_seed: true,
});

Tab

await tab.goto(url, "load", 30);
await tab.find(selector, 5);
await tab.find_all(selector);
await tab.wait_for_selector(selector, false, false, 30);
await tab.wait_for_function(expression, 30);
await tab.race([".success"], ["window.done === true"], false, 30);
await tab.evaluate("document.title");
await tab.screenshot("shot.png");
await tab.content();
await tab.sleep(2);

Element

await element.click();
await element.type("hello");
await element.text();
await element.inner_html();
await element.get_attribute("href");
await element.set_attribute("data-test", "1");
await element.is_visible();

Notes

• The port currently mirrors the real implemented behavior in the Python package.
• The dedicated fingerprint and runtime stealth modules were removed from the Node package because enabling them increased Cloudflare detection in testing.

Publish

Build the package first:

pnpm build

Inspect the publishable tarball locally:

pnpm pack

Log into npm:

npm login
npm whoami

Publish the scoped package publicly:

pnpm publish --access public --no-git-checks