Playwright + Claude Code : Write, Run and Fix Tests with AI Agents.

Claude Code + Playwright's AI agents (Planner, Generator, Healer) build a full Playwright test suite. Then I break the app 3 times to see if the AI fixes the right thing. 📢 Kane CLI by TestMu AI is a natural-language command line tool for testing: describe the test in natural language and run it from your terminal, your CI, or your own AI agent. Try Kane CLI: https://www.testmuai.com/kane-cli/?utm_source=youtube&utm_medium=organic&utm_campaign=playwright_claude_code_course What you'll see: 🔹 Claude Code sets up a Playwright workspace and writes the first test with role/label locators 🔹 Playwright Planner + Generator agents turn a requirements doc into a test suite 🔹 3 deliberate breaks: a renamed button, a calculator NaN bug, hidden button text 🔹 Where the Healer agent got it right, and where it fell short 🔹 The same jobs in Kane CLI: test.md natural-language tests, requirement-to-test coverage, failure classification, shareable evidence, export to Playwright code 🔹 API tests + a GitHub Actions workflow 🔹 What you still own: reviewing assertions, healer diffs and screenshot baselines 🚀 Install Kane CLI: npm install -g @testmuai/kane-cli 🔗 Sign up for TestMu AI: https://www.testmuai.com/ 📘 Kane CLI docs: https://www.testmuai.com/docs/kane-cli-use-cases 📗 Kane CLI cookbook: https://www.testmuai.com/kane-cli-cookbook/ ▶️ Related: How to Run Playwright Tests with Agent Skills Using Claude Code [LINK] Chapters 00:00 Playwright + Claude Code: what we're building 00:30 Setting up the Playwright workspace with Claude Code 00:58 First Playwright test: open account flow 01:27 Kane CLI: the same flow as a natural-language test.md 01:48 Playwright Planner agent: test plan from a requirements doc 02:08 Kane CLI: requirements to test coverage map 02:22 Playwright Generator agent vs Kane CLI generate 02:51 Debugging: Playwright UI mode vs Kane CLI evidence link 03:10 Break #1: renamed button (Healer agent vs Kane CLI) 03:38 Break #2: calculator bug (will the Healer hide it?) 04:12 Break #3: hidden button (screenshot test vs Kane CLI) 04:34 API tests with Claude Code 04:53 GitHub Actions CI + Kane CLI test run 05:07 Export Kane CLI tests to Playwright code 05:16 Verdict: what AI did vs what you still own #Playwright #ClaudeCode #TestAutomation #KaneCLI #TestMuAI