Puppeteer

Puppeteer connects to rayobrowse via puppeteer.connect() with a browserWSEndpoint. First call the HTTP /connect endpoint to get a CDP WebSocket URL.

Using /connect (no SDK)

const puppeteer = require('puppeteer-core');

(async () => {
  const resp = await fetch(
    'http://localhost:9222/connect?headless=false&os=windows'
  );
  const cdpUrl = (await resp.text()).trim();
  const browser = await puppeteer.connect({ browserWSEndpoint: cdpUrl });

  const page = (await browser.pages())[0] || await browser.newPage();
  await page.goto('https://example.com', { waitUntil: 'domcontentloaded' });
  console.log(await page.title());

  await browser.disconnect();
})();

Use puppeteer-core instead of puppeteer since the browser is provided by rayobrowse. No bundled Chromium download is needed.

Using the API

For more control, store the returned session ID for logs or reconnection:

const puppeteer = require('puppeteer-core');

const ENDPOINT = process.env.RAYOBYTE_ENDPOINT || 'http://localhost:9222';

async function createBrowser() {
  const resp = await fetch(`${ENDPOINT}/connect?headless=false&os=windows`);
  return {
    cdpUrl: (await resp.text()).trim(),
    sessionId: resp.headers.get('x-session-id'),
  };
}

const { cdpUrl, sessionId } = await createBrowser();
const browser = await puppeteer.connect({ browserWSEndpoint: cdpUrl });
console.log(`Session: ${sessionId}`);
// Automate pages...
await browser.disconnect();

With proxy

const resp = await fetch(
  'http://localhost:9222/connect?headless=true&os=windows&proxy=http://user:pass@host:port'
);
const browser = await puppeteer.connect({
  browserWSEndpoint: (await resp.text()).trim(),
});