Command line tools

Playwright comes with the command line tools.

Usage

playwright

Install browsers

Playwright can install supported browsers.

# Running without arguments will install default browsers
playwright install

You can also install specific browsers by providing an argument:

# Install WebKit
playwright install webkit

See all supported browsers:

playwright install --help

Install system dependencies

System dependencies can get installed automatically. This is useful for CI environments.

# See command help
playwright install-deps

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

playwright install-deps chromium

It's also possible to combine install-deps with install and install by that the browsers and OS dependencies with a single command. This would do both for Chromium, but you can also leave it out.

playwright install --with-deps chromium

Generate code

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.

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.

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 browser_context.route()), it is possible to call page.pause() that will open a separate window with codegen controls.

Sync

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()

Async

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())

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

Emulate color scheme and viewport size

# 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
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.$$(selector)

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

playwright.inspect(selector)

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

playwright.locator(selector)

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

playwright.selector(element)

Generates selector for the given element.

Take screenshot

# See command help
playwright screenshot --help

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

Generate PDF

PDF generation only works in Headless Chromium.

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