Playwright 1.59: Every New Feature Explained — The Most Important Release for AI-Era Testing

Playwright 1.59 is not just another release — it is a shift toward AI-driven testing. From the Screencast API to Browser.bind() for multi-client sessions, every major feature enables agentic testing workflows.

🤖 Learning AI-powered testing? Go hands-on with LLM, RAG, and AI-agent testing in the AI-Powered Testing Mastery course at The Testing Academy.

Screencast API: Video Recording for AI Agents

Record video, stream live frames for AI analysis, add chapters, and annotate actions programmatically. CI bots can now record visual walkthroughs that AI agents analyze for regression.

Browser.bind(): Multi-Client Sessions

One live browser with multiple clients connected simultaneously. This is game-changing for agentic and collaborative debugging — multiple AI agents can observe and interact with the same browser session.

Built-in Dashboard

playwright cs show — see all bound browsers, live status, and open DevTools directly from the browser. No more blind debugging in CI.

CLI Debugger for Agents

npx playwright test --debug=cli

Perfect for agentic CI pipelines where no GUI is available. Agents can step through tests, inspect state, and make decisions without a display.

await using Syntax: Auto-Cleanup

test('auto-cleanup example', async ({ browser }) => {
  await using page = await browser.newPage();
  // page is automatically closed when scope exits
  // No more forgotten cleanup in finally blocks
});

Other Highlights

• page.pickLocator() — interactive locator picker from the CLI
• locator.normalized() — whitespace-normalized text matching
• browserContext.setStorageState() — set auth state without file I/O
• New trace mode: retain-on-failure-and-retries

Breaking Change: macOS 14 WebKit Dropped

WebKit support now requires macOS 15+. If your CI runs WebKit tests on macOS 14, you need to upgrade before updating to Playwright 1.59.