Capture screenshots and generate PDFs from a page inside a Browserbase session

Documented steps

1. Connect to a running session and obtain a Playwright or Puppeteer page object.
2. For a screenshot, open a CDP session: Playwright const client = await context.newCDPSession(page); Puppeteer const client = await page.createCDPSession().
3. Call client.send("Page.captureScreenshot", { format: "jpeg", quality: 80, captureBeyondViewport: true }) and read the base64 data field from the result. Reference: https://docs.browserbase.com/platform/browser/files/screenshots
4. Persist it: Node fs.writeFileSync(path, Buffer.from(data, "base64")); Python open(path, "wb").write(base64.b64decode(screenshot_data)).
5. With Selenium no CDP is needed - call driver.save_screenshot("screenshot.png") directly.
6. For a PDF of the current page use Playwright's built-in generator: await page.pdf({ path: "webpage.pdf", format: "A4" }). Reference: https://docs.browserbase.com/platform/browser/files/pdfs
7. To capture an existing PDF at a URL instead, configure Browser.setDownloadBehavior via CDP ({ behavior: "allow", downloadPath: "downloads", eventsEnabled: true }) before navigating - navigating to a PDF URL auto-downloads it and cancels the navigation.
8. Retrieve that downloaded PDF via GET https://api.browserbase.com/v1/downloads?sessionId={sessionId}, then GET /v1/downloads/{id} with Accept: application/octet-stream.
9. If you want PDFs rendered in-browser rather than downloaded, set browserSettings.enablePdfViewer: true at session creation - but note this breaks the download-and-fetch flow.
10. Wait for the page to settle (network idle or an explicit delay) before capturing, since the documented Python examples insert 3-5 second delays to let generation complete.