Yesterday you wrote pointers by hand; today you install the tool that uses them, and you get your first green run before this email is old. Choosing a test automation tool is one of the bigger decisions a QA team makes, and people agonise over it. The honest answer is shorter than the agonising: the frameworks are all capable, and the tests you write matter far more than the logo on them.
The three you will hear about
- Playwright. Microsoft’s framework. Modern, fast, runs with a visible browser or invisibly in the background, and handles multiple tabs, flows that cross sites, and phone-sized screens without a fight. Its failure traces are the best in the business. This is what I reach for first on new web work, and it is what this course uses.
- Cypress. Developer-friendly, lovely local experience, superb debugging. It runs inside the browser, which historically limited multi-tab and cross-site flows, though newer versions have loosened that. I have shipped a lot of real work with it, and its intercept-and-wait pattern is one we will borrow on Day 9 no matter which tool you use.
- Selenium. The elder statesman. Widest language and browser coverage, huge community. Java shops with existing frameworks still get real value; I have shipped plenty with Selenium and Java too. For a first framework in 2026, I would not start here.
For a native mobile app you would reach for Appium instead, and that is a different course.
Pick one, learn it properly, ship a small suite, and ignore the speed benchmarks. Your bottleneck is test design, not framework speed.
Install Playwright, right now
You need Node.js (free from nodejs.org, click through the installer). You also need the terminal: the app where you type commands instead of clicking, called Terminal on a Mac and PowerShell on Windows. It comes with your computer; open it now. Three words before you type: mkdir makes a folder, cd steps into it, and npm (which arrived with Node) installs tools, while npx runs them. Now, in your terminal:
mkdir tumble-kitchen-tests
cd tumble-kitchen-tests
npm init playwright@latest
Accept the defaults. It scaffolds a small project with an example test. Run it:
npx playwright test
Green passes against a demo site, on your machine. Now the two views worth knowing:
npx playwright show-report
npx playwright test --ui
The first opens the HTML report: every test, every step, every timing. The second opens UI mode, where you can run a test and literally watch it step through the page. Ten minutes in there and automation stops being magic.
Point it at the store
One more tool: a code editor. Install VS Code (free, from code.visualstudio.com) and open your tumble-kitchen-tests folder in it. Do not write code in a word processor; it quietly swaps quote characters and breaks things. In VS Code, create a new file, tests/tumble-kitchen.spec.js, and type this in (typing beats pasting for learning, and it is short):
import { test, expect } from '@playwright/test';
test('verify that the practice store loads its catalogue', async ({ page }) => {
await page.goto('https://juliapottinger.com/practice/tumble-kitchen/');
await expect(page.getByRole('heading', { level: 1 })).toContainText('Tumble Kitchen');
await expect(page.getByTestId('product-card').first()).toBeVisible();
});
Run npx playwright test tumble-kitchen and you have a passing test against a real page on the internet, written by you. Notice the name: it is a sentence stating what it verifies. Every test you write from now on gets a name like that.
Do this today
The install, the green run, and the store test above, at a keyboard. Then open UI mode and watch your own test execute. Tomorrow we slow down for three short days of just enough JavaScript, so every line you typed today reads like a sentence instead of a spell.