Automate Web Application Testing with Playwright

Web Application Testing is a skill for automating local web apps with native Python Playwright scripts, built around a bundled scripts/with_server.py helper that manages one or more server processes' lifecycles so the automation script itself only needs to contain Playwright logic. It runs on a decision tree: for static HTML, read the file directly to identify selectors before writing a script; for a dynamic webapp with no server running, invoke the helper to start it; for a dynamic webapp already running, use a reconnaissance-then-action pattern - navigate and wait for networkidle, screenshot or inspect the DOM, identify selectors from the rendered state, then act on them.

Use it whenever a task calls for browser-driving a local web app rather than reading its raw source - clicking through a UI, verifying rendered content, or capturing console output. It explicitly favors treating the bundled scripts as black boxes: run them with --help first, and don't read their source unless a genuinely customized solution turns out to be necessary, since the scripts can be large and reading them pollutes the context window unnecessarily. The single most common pitfall it calls out: inspecting the DOM before page.wait_for_load_state('networkidle') on a dynamic app, which reads a page before its JavaScript has finished executing.

Input: a target local URL or HTML file, plus, for dynamic apps, the server start command(s) and port(s). Output: a Playwright automation script using sync_playwright(), launching Chromium in headless mode, that navigates, waits for networkidle, and then performs discovered actions using descriptive selectors (text=, role=, CSS, or IDs) with explicit waits (wait_for_selector or wait_for_timeout), always closing the browser when done.

python scripts/with_server.py --server "npm run dev" --port 5173 -- python your_automation.py

For multi-service apps, the same helper accepts multiple --server/--port pairs, for example a backend and a frontend, before the automation command. Bundled example scripts cover common patterns directly: element_discovery.py for finding buttons, links, and inputs on a page, static_html_automation.py for file:// URL testing, and console_logging.py for capturing console output during a run.

Developers who need to drive or verify a local web app's UI programmatically - checking that a page renders correctly, that interactive elements work, or that console output is clean - without manually reimplementing server-lifecycle management or DOM-discovery logic each time. The reconnaissance step itself is intentionally simple: a full-page screenshot, a call to page.content() for the raw rendered HTML, and page.locator('button').all() to enumerate interactive elements, all before any selector is written into the final script.