Supercharged Testing: AI-Powered Workflows with Playwright + MCP
summary
Demonstrates generating end to end tests by wiring a local LLM to the Playwright MCP Test Server and running three specialized agents planner/generator/healer agents that produce a test plan (example gave 10 suites and 45 cases), generate code that verifies elements before writing, and auto‑repair failing tests (one run reduces 23 failing to 16). Uses accessible tree snapshots not vision, logs in via project dependencies so credentials remain hidden, and the full flow for a small app takes about an hour to produce usable tests.
AI-written summary · about
chapters (23)
- 0:00Introduction
- 1:13Speaker Debbie O'Brien
- 3:42AI testing importance
- 5:26Playwright benefits
- 6:29Model Context Protocol
- 7:34Playwright agents concepts
- 9:17Bug discovery with AI
- 11:01AI test issues
- 12:44Debugging in Playwright
- 14:49Playwright agents intro
- 17:16MCP server comparison
- 19:43VS Code agents usage
- 21:29Planner agent demo
- 24:17Secure login dependencies
- 26:43Test plan generation
- 27:45Generator agent demo
- 29:47Healer agent demo
- 33:18Efficient testing best practices
- 35:23AI agents future
- 37:48Playwright upcoming features
- 40:40Large test suite management
- 42:46Call to action
- 47:20Closing & Q&A
related talks
14mAutonomous Testing Agents with LLMs in React
49mReal World Lessons from Scaling Test Automation
45mAI-Powered Debugging & Browser Automation with Playwright MCP
01h 31mWorkshop: Advanced Web Testing with Playwright and AI
39mHow to Use AI with Playwright for UI & API Testing
20mBeyond Code Coverage: Functionality Testing with Playwright MCP
25mAI-Powered Debugging & Browser Automation with Playwright MCP
25mThe 3 Pillars of Autonomy