@playwright-java

Scaffold, write, debug, and enhance enterprise-grade Playwright E2E tests in Java using Page Object Model, JUnit 5, Allure reporting, and parallel execution.

Supporting material

The bundled files are configuration, page objects, assertions, fixtures, BaseTest.java and BasePage.java. Copy/adapt them into the target project with their dependencies; src/test/resources/junit-platform.properties below is a project file to create, not a bundled file.

Playwright Java – Advanced Test Automation

Overview

This skill supplies Playwright Java test patterns requiring project-specific compilation and execution. It enforces the Page Object Model (POM), strict locator strategies, per-thread Playwright ownership for parallel execution, and full Allure reporting integration. Targets Java 17+ and Playwright 1.44+.

Supporting reference files are available for deeper topics:

TopicFile
Maven POM, ConfigReader, Docker/CI setupconfig.md
Component pattern, dropdowns, uploads, waitspage-objects.md
Full assertion API, soft assertions, visual testingassertions.md
Fixtures, test data factory, auth state, retryfixtures.md
Drop-in base class templatesBaseTest.java, BasePage.java

When to Use This Skill

  • Use when scaffolding a new Playwright Java project from scratch
  • Use when writing Page Object classes or JUnit 5 test classes
  • Use when the user asks about cross-browser testing, parallel execution, or Allure reports
  • Use when fixing flaky tests or replacing Thread.sleep() with proper waits
  • Use when setting up Playwright in CI/CD pipelines (GitHub Actions, Jenkins, Docker)
  • Use when combining API calls and UI assertions in a single test (hybrid testing)
  • Use when the user mentions "POM pattern", "BrowserContext", "Playwright fixtures", or "traces"

How It Works

Step 1: Decide the Approach

Use this matrix to pick the right pattern before writing any code:

User RequestApproach
New project from scratchFull scaffold — see config.md
Single feature testPOM page class + JUnit5 test class
API + UI hybridAPIRequestContext alongside Page
Cross-browser@MethodSource parameterized over browser names
Flaky test fixReplace sleep with waitFor / waitForResponse
CI integrationplaywright install --with-deps in pipeline
Parallel executionjunit-platform.properties + ThreadLocal
Rich reportingAllure + Playwright trace + video recording

Step 2: Scaffold the Project Structure

Always use this layout when creating a new project:

src/
├── test/
│   ├── java/com/company/tests/
│   │   ├── base/
│   │   │   ├── BaseTest.java        ← BaseTest.java
│   │   │   └── BasePage.java        ← BasePage.java
│   │   ├── pages/
│   │   │   └── LoginPage.java
│   │   ├── tests/
│   │   │   └── LoginTest.java
│   │   ├── utils/
│   │   │   ├── TestDataFactory.java
│   │   │   └── WaitUtils.java
│   │   └── config/
│   │       └── ConfigReader.java
│   └── resources/
│       ├── test.properties
│       ├── junit-platform.properties
│       └── testdata/users.json
pom.xml

Step 3: Set Up Thread-Safe BaseTest

Use the bundled BaseTest.java, with the matching ConfigReader from config.md; it registers resources before fallible tracing/page creation and removes thread-local values during teardown.


Step 4: Build Page Object Classes

public class LoginPage extends BasePage {

    // Declare ALL locators as fields — never inline in action methods
    private final Locator emailInput;
    private final Locator passwordInput;
    private final Locator loginButton;
    private final Locator errorMessage;

    public LoginPage(Page page) {
        super(page);
        emailInput    = page.getByLabel("Email address");
        passwordInput = page.getByLabel("Password");
        loginButton   = page.getByRole(AriaRole.BUTTON,
                            new Page.GetByRoleOptions().setName("Sign in"));
        errorMessage  = page.getByTestId("login-error");
    }

    @Override protected String getUrl() { return "/login"; }

    // Navigation methods return the next Page Object — enables fluent chaining
    public DashboardPage loginAs(String email, String password) {
        fill(emailInput, email);
        fill(passwordInput, password);
        clickAndWaitForNav(loginButton, "**/dashboard");
        return new DashboardPage(page);
    }

    public LoginPage loginExpectingError(String email, String password) {
        fill(emailInput, email);
        fill(passwordInput, password);
        loginButton.click();
        errorMessage.waitFor();
        return this;
    }

    public String getErrorMessage() { return errorMessage.textContent(); }
}

Step 5: Write Tests with Allure Annotations

class LoginTest extends BaseTest {

    private LoginPage loginPage;

    @BeforeEach
    void openLoginPage() {
        loginPage = new LoginPage(page());
        loginPage.navigate();
    }

    @Test
    @Severity(SeverityLevel.BLOCKER)
    @DisplayName("Valid credentials redirect to dashboard")
    void shouldLoginWithValidCredentials() {
        User user = TestDataFactory.getDefaultUser();
        DashboardPage dash = loginPage.loginAs(user.email(), user.password());

        assertThat(page()).hasURL(Pattern.compile(".*/dashboard"));
        assertThat(dash.getWelcomeBanner()).containsText("Welcome, " + user.firstName());
    }

    @Test
    void shouldShowErrorOnInvalidCredentials() {
        loginPage.loginExpectingError("bad@test.com", "wrongpass");

        SoftAssertions softly = new SoftAssertions();
        softly.assertThat(loginPage.getErrorMessage()).contains("Invalid email or password");
        assertThat(page()).hasURL(Pattern.compile(".*/login"));
        softly.assertAll();
    }

    @ParameterizedTest
    @MethodSource("provideInvalidCredentials")
    void shouldRejectInvalidCredentials(String email, String password, String expectedError) {
        loginPage.loginExpectingError(email, password);
        org.assertj.core.api.Assertions.assertThat(loginPage.getErrorMessage()).contains(expectedError);
    }

    static Stream<Arguments> provideInvalidCredentials() {
        return Stream.of(
            Arguments.of("", "password123", "Email is required"),
            Arguments.of("user@test.com", "", "Password is required"),
            Arguments.of("notanemail", "pass", "Invalid email format")
        );
    }
}

Examples

Example 1: API + UI Hybrid Test

@Test
void shouldDisplayNewlyCreatedOrder() {
    // Arrange via API — faster than navigating through UI
    APIRequestContext api = page().context().request();
    APIResponse response = api.post("/api/orders",
        RequestOptions.create()
            .setHeader("Authorization", "Bearer " + authToken)
            .setData(Map.of("productId", "SKU-001", "quantity", 2)));
    assertThat(response).isOK();

    String orderId = new JsonParser().parse(response.text())
        .getAsJsonObject().get("id").getAsString();

    OrdersPage orders = new OrdersPage(page());
    orders.navigate();
    assertThat(orders.getOrderRowById(orderId)).isVisible();
}

Example 2: Network Mocking

@Test
void shouldHandleApiFailureGracefully() {
    page().route("**/api/products", route -> route.fulfill(
        new Route.FulfillOptions()
            .setStatus(503)
            .setBody("{\"error\":\"Service Unavailable\"}")
            .setContentType("application/json")));

    ProductsPage products = new ProductsPage(page());
    products.navigate();

    assertThat(products.getErrorBanner())
        .hasText("We're having trouble loading products. Please try again.");
}

Example 3: Cross-Browser Execution

Run separate Maven processes/jobs with -Dbrowser=chromium, firefox and webkit. Do not set a process-global system property inside a parameterized test: setup already launched the browser and parallel tests race on that property. Give each job a unique artifact name.

Example 4: Parallel Execution Config

# src/test/resources/junit-platform.properties
junit.jupiter.execution.parallel.enabled=true
junit.jupiter.execution.parallel.mode.default=concurrent
junit.jupiter.execution.parallel.config.strategy=fixed
junit.jupiter.execution.parallel.config.fixed.parallelism=4

Example 5: GitHub Actions CI Pipeline

- name: Install Playwright browsers
  run: mvn exec:java -e -Dexec.mainClass=com.microsoft.playwright.CLI -Dexec.args="install --with-deps"

- name: Run tests
  run: mvn test -Dbrowser=${{ matrix.browser }} -Dheadless=true

- name: Upload traces on failure
  uses: actions/upload-artifact@v4
  if: failure()
  with:
    name: playwright-traces-${{ matrix.browser }}
    path: target/traces/

- name: Upload Allure results
  uses: actions/upload-artifact@v4
  if: always()
  with:
    name: allure-results-${{ matrix.browser }}
    path: target/allure-results/

Use an authorized test origin and disposable test data/accounts. Order/API/UI actions can charge, send messages or delete records; block production integrations and clean up created records. Storage-state files contain credentials, and traces/screenshots/videos can contain secrets or personal data: restrict access, ignore in git, redact and bound retention before uploading. Pinned versions in the configuration are historical examples, not current recommendations; match Java dependency and browser/Docker versions.

Best Practices

  • ✅ Use ThreadLocal<Page> for every parallel-safe test suite
  • ✅ Declare all Locator fields at the top of the Page Object class
  • ✅ Return the next Page Object from navigation methods (fluent chaining)
  • ✅ Use assertThat(locator) — it auto-retries until timeout
  • ✅ Use getByRole, getByLabel, getByTestId as first-choice locators
  • ✅ Start tracing in @BeforeEach and stop with a file path in @AfterEach
  • ✅ Use SoftAssertions when validating multiple fields on a single page
  • ✅ Set up saved auth state (storageState) to skip login across test classes
  • ❌ Never use Thread.sleep() — replace with waitFor() or waitForResponse()
  • ❌ Never hardcode base URLs — always use ConfigReader.getBaseUrl()
  • ❌ Never create a Playwright instance inside a Page Object
  • ❌ Never use XPath for dynamic or frequently changing elements

Common Pitfalls

  • Problem: Tests fail randomly in parallel mode Solution: Ensure every test creates its own Playwright → Browser → BrowserContext → Page chain via ThreadLocal. Never share a Page across threads.

  • Problem: assertThat(locator).isVisible() times out even when the element appears Solution: Inspect readiness/locator first; set the assertion-specific options timeout if justified. Assertions have a separate default (5 seconds); context action timeout does not change it.

  • Problem: Thread.sleep(2000) was added but tests are still flaky Solution: Replace with page.waitForResponse("**/api/endpoint", () -> action()) or assertThat(locator).hasText("Done") which polls automatically.

  • Problem: Playwright trace zip is empty or missing Solution: Ensure tracing().start() is called before test actions and tracing().stop() is in @AfterEach — not @AfterAll.

  • Problem: Allure report is blank or missing steps Solution: Add the AspectJ agent to maven-surefire-plugin <argLine> in pom.xml — see config.md for the exact snippet.

  • Problem: storageState auth file is stale and tests redirect to login Solution: Re-run AuthSetup to regenerate target/auth/user-state.json before the suite, or add a @BeforeAll that conditionally refreshes it.


  • @rest-assured-java — Use for pure API test suites without any UI interaction
  • @selenium-java — Legacy alternative; prefer Playwright for all new projects
  • @allure-reporting — Deep-dive into Allure annotations, categories, and history trends
  • @testcontainers-java — Use alongside this skill when tests need a live database or service
  • @github-actions-ci — For building complete multi-browser matrix CI pipelines

Limitations

  • Use this skill only when the task clearly matches the scope described above.
  • Do not treat the output as a substitute for environment-specific validation, testing, or expert review.
  • Stop and ask for clarification if required inputs, permissions, safety boundaries, or success criteria are missing.