How to use Playwright — from install to your first test run

If you follow this from the top, learning how to use Playwright takes about 10 minutes from install to first test run. One command installs it, an example test comes with it, and one command runs it. This post walks that 10 minutes in order, and then covers what is still left once you get there.

How to use Playwright — what you need first

Node.js. That is the whole list. Browsers come down during install.

You do not need an app to test yet either. The installer writes an example test that runs against the Playwright site, so you can run it as-is.

Install, configure, run, review — four steps to your first Playwright run

How to install Playwright

From your project folder:

npm init playwright@latest

For other package managers, yarn create playwright and pnpm create playwright.

Per the official documentation, the installer asks four questions.

The browser download is the slow part. It is a sizeable payload the first time.

What gets created

When it finishes, the layout looks like this.

playwright.config.ts    configuration
package.json
tests/
  example.spec.ts       example test

playwright.config.ts holds which browsers to run, whether to retry on failure, whether to record traces. You do not need to touch it at first.

tests/example.spec.ts is the example. Reading that one file tells you most of what a test looks like.

The installer writes a config file, an example test and optionally a CI workflow

Running your first test

One command.

npx playwright test

A browser opens without a visible window, the example runs, and the terminal reports pass or fail. The default config runs against Chromium, Firefox and WebKit, so two written tests show up as six results.

To watch it happen:

npx playwright test --headed

Reading the results

When terminal output is not enough:

npx playwright show-report

A report opens in the browser with pass, fail and duration per test. Click a failure and you can see where it stopped.

Keep readingWhat is Playwright trace viewer? What is left when a test failsplaywright trace viewerPlaywright5 min

Pointing it at your own app

Once the example makes sense, swap in your own screen. A test looks roughly like this.

import { test, expect } from '@playwright/test';

test('signing in lands on the dashboard', async ({ page }) => {
  await page.goto('http://localhost:3000/login');

  await page.getByLabel('Email').fill('me@example.com');
  await page.getByLabel('Password').fill('password');
  await page.getByRole('button', { name: 'Sign in' }).click();

  await expect(page.getByRole('heading', { name: 'Dashboard' })).toBeVisible();
});

It reads as the same sequence a person would follow. Open the page, fill the fields, click the button, confirm the expected screen appeared.

getByLabel and getByRole are how the test points at elements. Which of those you choose determines how long the test survives.

Keep readingWhat are Playwright locators? Writing tests that survive a redesignplaywright locatorsPlaywright6 min

If writing them by hand is slow, you can have your clicks recorded into code instead. The codegen post covers that route.

npx playwright codegen http://localhost:3000

What is left once you get here

Install and first run take 10 minutes. The rest does not.

First, deciding what to test. A tool checks only what you told it to. In a product with 20 screens, which flows matter and where the bar for "passing" sits is yours to set.

Second, updating tests as the UI changes. Change a feature and the test changes with it, and that work keeps coming back.

Third, deciding when they run. Anything triggered by hand gets forgotten. They have to run on every deploy to mean anything.

All three need someone who can read and edit code. Learning how to use Playwright and keeping a suite alive are different jobs.

Specnote takes the part after that

If you can read and edit code, writing tests directly in Playwright is more precise and more flexible. Start with the steps above.

Specnote is for the case where nobody has the time or the person for that work. AI lists what it built, you review and approve the list, and only the approved scope runs in a real browser. Nothing to install, no code to write.

Commands and setup steps in this post were verified against the Playwright installation docs on August 21, 2026.