Cypress vs Playwright vs Selenium: choosing a test automation framework

When a team starts with end-to-end test automation, the framework question comes up quickly: Cypress, Playwright or Selenium? We use all three on client projects. Each is a sound choice in the right situation, and each can be painful in the wrong one.

Playwright

Playwright is an open-source framework from Microsoft. One API drives Chromium, Firefox and WebKit (the engine behind Safari), and tests can be written in JavaScript or TypeScript, Python, Java or .NET.

• Waits automatically for elements to be ready, which removes most flaky sleeps
• Handles multiple tabs, multiple users and different domains in one test
• Runs tests in parallel out of the box and records traces for debugging failures
• Includes API testing, so UI and API checks can live in one suite

It is our default for new projects, especially when Safari coverage matters or when a test has to switch between an admin and a customer session.

Cypress

Cypress runs tests inside the browser, next to the application. Its interactive runner, with time-travel snapshots of every step, is one of the best debugging experiences in testing, and front-end developers usually pick it up quickly.

• Tests are written in JavaScript or TypeScript only
• Excellent developer experience and readable test code
• Strong support for stubbing network requests and component testing
• Multi-tab and cross-domain flows need workarounds such as cy.origin

Cypress is a good fit for JavaScript teams testing a single web application who want developers to write and maintain tests themselves.

Selenium

Selenium is the longest-established option and is built on the W3C WebDriver standard. It supports the widest range of languages and browsers, and Selenium Grid lets large suites run across many machines. Appium, the common choice for mobile app automation, uses the same WebDriver approach.

• Bindings for Java, Python, C#, JavaScript, Ruby and more
• Huge ecosystem, community and hiring pool
• More setup work, and waits must be handled carefully to avoid flaky tests
• A natural choice when an organisation already has a Selenium suite

How we choose

• New web project, mixed browsers, complex flows: Playwright
• JavaScript team that wants developers owning tests: Cypress
• Existing Selenium suite, Java or C# QA team, or a large grid: Selenium
• Mobile apps: Appium, alongside one of the above for the web

The best framework is the one your team will keep running on every change.

Whatever you pick

Most failing automation projects fail for reasons that have nothing to do with the framework: tests that depend on shared data, selectors tied to styling, and suites that only run the night before a release. Stable test data, test ids in the markup and running the suite on every pull request matter more than the tool on the box.