ZeekrBaha/playwright-qa-agent-generator-healer

veriplay is an autonomous QA agent that explores a web app in a real Chromium session and emits a deterministic Playwright test suite that has been observed passing (specs, Page Object Model, and an axe-core a11y check). It's aimed at QA engineers and developers who want generated, verifiable Playwright tests and automated maintenance. Key features: - Three-stage LLM-driven pipeline (Planner → Explorer → Critic) that plans scenarios, drives the browser via typed tools, and validates traces. - Selector cascade (getByRole → getByLabel → placeholder → testId → CSS) and per-host memory to prefer robust locators. - Transcriber produces tests/pages a11y checks and a JSON report; outputs a runnable Playwright project. - heal command parses Playwright failure reports, distinguishes selector misses, asks an LLM to propose and verify replacement selectors, and rewrites specs. - Integrates Playwright, OpenAI, and axe-core; examples included (saucedemo).

AI-related keywords found in this repo's description, topics, or README summary — grouped by category. Each badge links to the corresponding ranking detail page.