microsoft/playwright-component-testing
Set up component testing with Playwright using a story gallery — scaffold stories and a gallery dev page driven by the built-in mount fixture, no dedicated component-testing runtime. Use when asked to test React or Vue components in isolation with Playwright, or to migrate off @playwright/experimental-ct-react / -vue.
npx skills add https://github.com/microsoft/playwright --skill playwright-component-testing
Test components with regular Playwright e2e tests against a small story gallery page hosted by the app's own dev server. No extra test runner, bundler integration or npm packages are required.
*.story.tsx (or .ts/.jsx/.js/.vue) files; each named export is one story.references/gallery-spec.md: it exposes window.mount(params) / window.unmount() that render a story — resolved from your story files (e.g. with import.meta.glob) — into #root. It is framework-specific and yours to own — there is no template to copy for it.mount(storyId, props?) fixture (from @playwright/test) drives the gallery's window.mount and returns a Locator for the gallery root (#root). Scope the queries from there — component.getByRole('button').click(), not component.click(). Nothing to scaffold for it.Everything the component needs must be set up *inside the story* (it runs in the browser); everything the test asserts must be observable *through the page* (DOM, URL, network). Where the component takes callbacks, the story creates the state, provides the callbacks and records the state into a hidden form for the test to assert on. mount(id, props) passes plain serializable props to the story.
vite.config.*): the gallery is served by the existing dev server at /playwright/gallery/index.html — Vite serves any .html file under the project root, the app's plugins/aliases/CSS apply automatically, and vite build ignores it. No extra server needed.baseURL at it. Requires vite and the framework plugin as devDependencies.references/gallery-spec.md: a page at <project>/playwright/gallery/ that renders the requested story into #root. Start from the worked example in the spec and the framework notes in references/react.md / references/vue.md. Keep story discovery (import.meta.glob) and the framework mount here — this is the only framework-specific glue, so keep it small. Import the app's global CSS the same way the app's own entry does.playwright.config.ts: projects: [
{
name: 'components',
testDir: './tests/components',
use: { ...devices['Desktop Chrome'], baseURL: 'http://localhost:5173/playwright/gallery/index.html', serviceWorkers: 'block', reuseContext: true },
},
],
webServer: {
command: 'npm run dev', // or: npx vite --config playwright/vite.config.ts
url: 'http://localhost:5173/playwright/gallery/index.html', // standalone server: http://localhost:3100/playwright/gallery/index.html
reuseExistingServer: !process.env.CI,
},
Match the port to the dev server. mount navigates to baseURL, so set baseURL to the gallery's URL. serviceWorkers: 'block' keeps the app's own service worker from serving cached responses that would shadow your page.route() mocks. reuseContext: true reuses the browser context across tests in a worker (as the old component-testing runtime did) — a large speedup for component suites. If the config already has projects/webServer, merge instead of replacing.
templates/<react|vue>/Button.story.*.templates/react/button.spec.ts, importing test/expect from @playwright/test.npx playwright test --project=components. Open http://localhost:5173/playwright/gallery/index.html in a browser to eyeball all stories.src/ without the .story.* extension, plus the export name — src/components/Button.story.tsx export Primary → components/Button/Primary. Any unique suffix works too: mount('Button/Primary'). A .story.vue single-file component is one story, addressable by its path alone (its default export).Examples are React; the Vue equivalents differ only in story syntax.
The story owns the state and provides the callbacks. Where the component takes callbacks, create the state inside the story, wire the callbacks to it, and record the state into a hidden form next to the component. Tests perform operations and assert on the recorded values:
export const Stateful = () => {
const [expanded, setExpanded] = useState(false);
return <>
<Expandable expanded={expanded} setExpanded={setExpanded} title="Title">Details</Expandable>
<form hidden><input data-testid="expanded" readOnly value={String(expanded)} /></form>
</>;
};
test('click should expand', async ({ mount }) => {
const component = await mount('components/Expandable/Stateful');
await component.locator('.codicon-chevron-right').click();
await expect(component.getByTestId('expanded')).toHaveValue('true');
});
This keeps the whole scenario in the browser: no callback marshalling, the story doubles as documentation, and the recorded state is visible when eyeballing the gallery. Record each observed value in its own data-testid input (String(...) or JSON.stringify(...) for payloads) and assert with toHaveValue() — a web-first assertion that retries until the state lands. The negative direction works the same way: perform the operation, then assert the value did not change.
When a scenario is genuinely parametric (e.g. a boundary-value sweep), pass props as the second argument to mount; the gallery hands them to the story as its props. Keep props to plain serializable data — callbacks belong inside the story.
export const WithTitle = ({ title = 'Default' }: { title?: string }) =>
<Button title={title} />;
const component = await mount('components/Button/WithTitle', { title: 'Hello' });
mount is generic over the story: pass the story type as a template argument to type-check the props (and update()):
import type { WithTitle } from './Button.story';
const component = await mount<typeof WithTitle>('components/Button/WithTitle', { title: 'Hello' });
This works for React and Vue stories alike; Vue stories must additionally declare the props at runtime — see the Typed props sections in references/react.md / references/vue.md.
update()To test how a component reacts to a prop change without remounting (state preserved), call component.update(newProps) — it re-renders the same story with new props on the existing root:
const component = await mount('components/Counter/Default', { value: 1 });
await expect(component.getByTestId('value')).toHaveText('1');
await component.update({ value: 2 });
await expect(component.getByTestId('value')).toHaveText('2');
This requires the gallery to reuse its root/instance (references/gallery-spec.md); state survives as long as the story stays the same.
Each mount() navigates fresh, so tests are fully isolated and mounting several stories in one test is cheap:
await expect(await mount('Button/Primary')).toHaveScreenshot('primary.png');
await expect(await mount('Button/Disabled')).toHaveScreenshot('disabled.png');
For visual comparison, screenshot the returned root locator (as above), not the page, to avoid asserting on browser chrome.
Use page.route() as usual — register routes before mount(), since mounting navigates. serviceWorkers: 'block' (set in the config above) keeps the app's own service worker from serving cached responses that shadow the routes. Teams with MSW handler libraries can start the worker inside a story or decorator instead.
Open your gallery URL (baseURL) in a browser and call await window.mount({ story: 'components/Button/Primary' }) from the devtools console — that is exactly what the mount fixture does. An unknown story rejects window.mount, which surfaces as the test's mount() throwing with a real stack. To browse without the console, give your gallery an optional index page.
src layouts: change the glob and the id derivation in your gallery (references/gallery-spec.md) to match.decorator helper next to the gallery and wrap components in stories; see references/react.md / references/vue.md.references/gallery-spec.md — the gallery endpoint contract to implement (start here).references/react.md — React walkthrough: providers, StrictMode, CSS.references/vue.md — Vue walkthrough: .story.ts and .story.vue stories, plugins.references/migration.md — migrating off @playwright/experimental-ct-react / -vue.Take microsoft/playwright-component-testing from the repository into ~/.claude/skills for personal
use, or into .claude/skills inside a project.
The agent identifies a skill by the name field in its header. Two skills with the
same name cannot sit side by side — one of them will be ignored.
The instructions reference npx.
Without those the skill loads but fails at the first command.