playwright/docs/src/cli.md

id	title
cli	Command Line Interface

Playwright comes with the command line tools that run via npx or as a part of the npm scripts.

Usage

npx playwright --help

mvn exec:java -e -Dexec.mainClass=com.microsoft.playwright.CLI

playwright

# Running from `package.json` script
{
  "scripts": {
    "help": "playwright --help"
  }
}

Generate code

npx playwright codegen wikipedia.org

mvn exec:java -e -Dexec.mainClass=com.microsoft.playwright.CLI -Dexec.args="codegen wikipedia.org"

playwright codegen wikipedia.org

Run codegen and perform actions in the browser. Playwright CLI will generate JavaScript code for the user interactions. codegen will attempt to generate resilient text-based selectors.

Preserve authenticated state

Run codegen with --save-storage to save cookies and localStorage at the end. This is useful to separately record authentication step and reuse it later.

npx playwright codegen --save-storage=auth.json
# Perform authentication and exit.
# auth.json will contain the storage state.

mvn exec:java -e -Dexec.mainClass=com.microsoft.playwright.CLI -Dexec.args="codegen  --save-storage=auth.json"
# Perform authentication and exit.
# auth.json will contain the storage state.

playwright codegen --save-storage=auth.json
# Perform authentication and exit.
# auth.json will contain the storage state.

Run with --load-storage to consume previously loaded storage. This way, all cookies and localStorage will be restored, bringing most web apps to the authenticated state.

npx playwright open --load-storage=auth.json my.web.app
npx playwright codegen --load-storage=auth.json my.web.app
# Perform actions in authenticated state.

mvn exec:java -e -Dexec.mainClass=com.microsoft.playwright.CLI -Dexec.args="open --load-storage=auth.json my.web.app"
mvn exec:java -e -Dexec.mainClass=com.microsoft.playwright.CLI -Dexec.args="codegen --load-storage=auth.json my.web.app"
# Perform authentication and exit.
# auth.json will contain the storage state.

playwright open --load-storage=auth.json my.web.app
playwright codegen --load-storage=auth.json my.web.app
# Perform actions in authenticated state.

Codegen with custom setup

If you would like to use codegen in some non-standard setup (for example, use [method: BrowserContext.route]), it is possible to call [method: Page.pause] that will open a separate window with codegen controls.

const { chromium } = require('playwright');

(async () => {
  // Make sure to run headed.
  const browser = await chromium.launch({ headless: false });

  // Setup context however you like.
  const context = await browser.newContext({ /* pass any options */ });
  await context.route('**/*', route => route.continue());

  // Pause the page, and start recording manually.
  const page = await context.newPage();
  await page.pause();
})();

// FIXME
import com.microsoft.playwright.*;

public class Example {
  public static void main(String[] args) {
    try (Playwright playwright = Playwright.create()) {
      BrowserType chromium = playwright.chromium();
      // Make sure to run headed.
      Browser browser = chromium.launch(new BrowserType.LaunchOptions().setHeadless(false));
      // Setup context however you like.
      BrowserContext context = browser.newContext(/* pass any options */);
      context.route("**/*", route -> route.resume());
      // Pause the page, and start recording manually.
      Page page = context.newPage();
      page.pause();
    }
  }
}

import asyncio
from playwright.async_api import async_playwright

async def main():
    async with async_playwright() as p:
        # Make sure to run headed.
        browser = await p.chromium.launch(headless=False)

        # Setup context however you like.
        context = await browser.new_context() # Pass any options
        await context.route('**/*', lambda route: route.continue_())

        # Pause the page, and start recording manually.
        page = await context.new_page()
        await page.pause()

asyncio.run(main())

from playwright.sync_api import sync_playwright

with sync_playwright() as p:
    # Make sure to run headed.
    browser = p.chromium.launch(headless=False)

    # Setup context however you like.
    context = browser.new_context() # Pass any options
    context.route('**/*', lambda route: route.continue_())

    # Pause the page, and start recording manually.
    page = context.new_page()
    page.pause()

Open pages

With open, you can use Playwright bundled browsers to browse web pages. Playwright provides cross-platform WebKit builds that can be used to reproduce Safari rendering across Windows, Linux and macOS.

# Open page in Chromium
npx playwright open example.com

# Open page in Chromium
mvn exec:java -e -Dexec.mainClass=com.microsoft.playwright.CLI -Dexec.args="open example.com"

# Open page in Chromium
playwright open example.com

# Open page in WebKit
npx playwright wk example.com

# Open page in WebKit
mvn exec:java -e -Dexec.mainClass=com.microsoft.playwright.CLI -Dexec.args="wk example.com"

# Open page in WebKit
playwright wk example.com

Emulate devices

open can emulate mobile and tablet devices from the playwright.devices list.

# Emulate iPhone 11.
npx playwright open --device="iPhone 11" wikipedia.org

# Emulate iPhone 11.
mvn exec:java -e -Dexec.mainClass=com.microsoft.playwright.CLI -Dexec.args='open --device="iPhone 11" wikipedia.org'

# Emulate iPhone 11.
playwright open --device="iPhone 11" wikipedia.org

Emulate color scheme and viewport size

# Emulate screen size and color scheme.
npx playwright open --viewport-size=800,600 --color-scheme=dark twitter.com

# Emulate screen size and color scheme.
mvn exec:java -e -Dexec.mainClass=com.microsoft.playwright.CLI -Dexec.args="open --viewport-size=800,600 --color-scheme=dark twitter.com"

# Emulate screen size and color scheme.
playwright open --viewport-size=800,600 --color-scheme=dark twitter.com

Emulate geolocation, language and timezone

# Emulate timezone, language & location
# Once page opens, click the "my location" button to see geolocation in action
npx playwright open --timezone="Europe/Rome" --geolocation="41.890221,12.492348" --lang="it-IT" maps.google.com

# Emulate timezone, language & location
# Once page opens, click the "my location" button to see geolocation in action
mvn exec:java -e -Dexec.mainClass=com.microsoft.playwright.CLI -Dexec.args='open --timezone="Europe/Rome" --geolocation="41.890221,12.492348" --lang="it-IT" maps.google.com'

# Emulate timezone, language & location
# Once page opens, click the "my location" button to see geolocation in action
playwright open --timezone="Europe/Rome" --geolocation="41.890221,12.492348" --lang="it-IT" maps.google.com

Inspect selectors

During open or codegen, you can use following API inside the developer tools console of any browser.

playwright.$(selector)

Query Playwright selector, using the actual Playwright query engine, for example:

> playwright.$('.auth-form >> text=Log in');

<button>Log in</button>

playwright.$$(selector)

Same as playwright.$, but returns all matching elements.

> playwright.$$('li >> text=John')

> [<li>, <li>, <li>, <li>]

playwright.inspect(selector)

Reveal element in the Elements panel (if DevTools of the respective browser supports it).

> playwright.inspect('text=Log in')

playwright.selector(element)

Generates selector for the given element.

> playwright.selector($0)

"div[id="glow-ingress-block"] >> text=/.*Hello.*/"

Take screenshot

# See command help
npx playwright screenshot --help

# See command help
mvn exec:java -e -Dexec.mainClass=com.microsoft.playwright.CLI -Dexec.args="screenshot --help"

# See command help
playwright screenshot --help

# Wait 3 seconds before capturing a screenshot after page loads ('load' event fires)
npx playwright screenshot \
    --device="iPhone 11" \
    --color-scheme=dark \
    --wait-for-timeout=3000 \
    twitter.com twitter-iphone.png

# Wait 3 seconds before capturing a screenshot after page loads ('load' event fires)
mvn exec:java -e -Dexec.mainClass=com.microsoft.playwright.CLI -Dexec.args='screenshot --device="iPhone 11" --color-scheme=dark --wait-for-timeout=3000 twitter.com twitter-iphone.png'

# Wait 3 seconds before capturing a screenshot after page loads ('load' event fires)
playwright screenshot \
    --device="iPhone 11" \
    --color-scheme=dark \
    --wait-for-timeout=3000 \
    twitter.com twitter-iphone.png

# Capture a full page screenshot
npx playwright screenshot --full-page en.wikipedia.org wiki-full.png

# Capture a full page screenshot
mvn exec:java -e -Dexec.mainClass=com.microsoft.playwright.CLI -Dexec.args='screenshot --full-page en.wikipedia.org wiki-full.png'

# Capture a full page screenshot
playwright screenshot --full-page en.wikipedia.org wiki-full.png

Generate PDF

PDF generation only works in Headless Chromium.

# See command help
npx playwright pdf https://en.wikipedia.org/wiki/PDF wiki.pdf

# See command help
mvn exec:java -e -Dexec.mainClass=com.microsoft.playwright.CLI -Dexec.args="pdf https://en.wikipedia.org/wiki/PDF wiki.pdf"

# See command help
playwright pdf https://en.wikipedia.org/wiki/PDF wiki.pdf

Install system dependencies

Ubuntu 18.04 and Ubuntu 20.04 system dependencies can get installed automatically. This is useful for CI environments.

# See command help
npx playwright install-deps

# See command help
mvn exec:java -e -Dexec.mainClass=com.microsoft.playwright.CLI -Dexec.args="install-deps"

# See command help
playwright install-deps

You can also install the dependencies for a single browser only by passing it as an argument:

npx playwright install-deps chromium

mvn exec:java -e -Dexec.mainClass=com.microsoft.playwright.CLI -Dexec.args="install-deps chromium"

playwright install-deps chromium

Known limitations

Opening WebKit Web Inspector will disconnect Playwright from the browser. In such cases, code generation will stop.