Skip to main
Article
Overhead view of a cargo ship in water
loaded with multi-colored shipping containers.

Running Playwright Inside Docker Containers

Learn how to run Playwright in headed mode to interact with the browser’s user interface from outside Docker containers.

Update :

This article was updated to reflect changes to Playwright and Docker:

  • Add section about debugging an existing test suite.
  • Encourage usage of Playwright UI mode where available.
  • Use version-less Playwright Docker image to stay current.
  • Replace deprecated docker-compose commands with docker compose.

Playwright is a test runner that uses real browsers to test web applications (an alternative to tools like Selenium). By default, Playwright runs these browsers in headless mode, which means the pages are loaded and tested without opening the browser window. This is great when running entire test suites locally, and in CI where having a bunch of rapidly opening windows would be disconcerting. However, when it comes to writing or debugging individual tests, it is convenient to open the browser in headed mode to actually see the page being tested.

Playwright for JS includes UI mode to let you run end-to-end tests with a graphical explorer and debugger, allowing you to view, run, and debug tests in a real browser window. For other supported platforms (like Python), Playwright includes the Playwright Inspector to conveniently launch the browser in headed mode and a separate window to control test execution. The inspector is launched by setting PWDEBUG=1 before calling playwright, and is compatible with all browsers supported by Playwright (Safari, Chrome, and Firefox) in all major operating systems.

Additionally, Playwright also includes a playwright open <url> subcommand to quickly launch the inspector on any URL. Examples in this post use playwright open <url>, but they also apply to UI mode and the inspector.

For convenience, Playwright provides a Docker image, which includes all browsers pre-installed and configured so you can skip the dependency installation steps. But what happens if you try to run Playwright with a headed browser inside a Docker container – which normally doesn’t have a graphical user interface?

# We expect a browser window to open and load google.com
docker run --rm mcr.microsoft.com/playwright npx -y playwright open google.com
Looks like you launched a headed browser without having a XServer running.
Set either 'headless: true' or use 'xvfb-run <your-playwright-app>' before running Playwright.

<3 Playwright Team

No browser window is launched! Instead we get an error message about not “having a XServer running.” The definition and functionality of XServer are beyond the scope of this article, but without it we can’t interact with applications that require a user interface (like the browser). Here’s a more detailed explanation if you want to learn more.

Searching for this error on the web will return results explaining how to install and start XServer. That advice applies to non-containerized, Linux-based systems. If your host system is macOS or Windows you actually don’t want to do that. Instead we want the container to use the host XServer to launch Playwright inside the container, which requires two modifications to our docker command:

  1. Set the DISPLAY environment variable inside the container using the -e option
  2. Mount the XServer Unix socket inside the container using the -v option
docker run --rm \
-e DISPLAY=<host display> \
-v /tmp/.X11-unix:/tmp/.X11-unix \
mcr.microsoft.com/playwright npx -y playwright open google.com

The value of <host display> will depend on your host operating system, and you will need to ensure /tmp/.X11-unix is available for mounting. The following sections explain how to do this for Windows and macOS.

You might find it surprising (I certainly did) that Microsoft Windows has a native XServer even though it’s not a GNU/Linux system. It’s called WSLg, and it’s included as part of the Windows Subsystem for Linux (WSL). You most likely already have WSL and WSLg installed if you are running Docker Desktop in recent builds of Windows 10 and 11.

Let’s start by verifying that WSL and WSLg are installed and running. First, launch “WSL” from your Start Menu. A Linux terminal window should open (most likely a recent version of Ubuntu). In that window, verify that the directory /mnt/wslg/ exists and contains these files inside the Linux filesystem:

ls -a -w 1 /mnt/wslg

.
..
.X11-unix
PulseAudioRDPSink
PulseAudioRDPSource
PulseServer
distro
doc
dumps
pulseaudio.log
runtime-dir
stderr.log
versions.txt
weston.log
wlog.log

If you don’t see “WSL” in your Start Menu, or the ls command above fails with No such file or directory, then your system is missing WSL entirely or is running an old version. Visit the Microsoft Store to download an up-to-date version.

Once you are all set up, we can set DISPLAY=:0 as explained in the official guide:

docker run --rm \
-e DISPLAY=:0 \
-v /tmp/.X11-unix:/tmp/.X11-unix \
mcr.microsoft.com/playwright npx -y playwright open google.com

If all goes well, that should open two windows: a browser window with Google loaded, and a Playwright Inspector window. Closing both will also stop the container.

Apple’s operating system doesn’t include a built-in XServer, but we can use XQuartz to provide one:

  1. Install XQuartz: brew install --cask xquartz
  2. Open XQuartz, go to Preferences -> Security, and check “Allow connections from network clients”
  3. Restart your computer (restarting XQuartz might not be enough)
  4. Start XQuartz with xhost +localhost
  5. Open Docker Desktop and edit settings to give access to /tmp/.X11-unix in Preferences -> Resources -> File sharing

Once XQuartz is running with the right permissions, you can populate the environment variable and socket:

docker run --rm \
-e DISPLAY=host.docker.internal:0 \
-v /tmp/.X11-unix:/tmp/.X11-unix \
mcr.microsoft.com/playwright npx -y playwright open google.com

If all goes well, that should open two windows: a browser window with Google loaded, and a Playwright Inspector window. Closing both will also stop the container.

To avoid having to specify the flags every time, you can use a docker-compose.yml file to set the environment variable and mount the socket.

services:
  web:
    image: mcr.microsoft.com/playwright
    environment:
      - DISPLAY=... # Replace this line with the appropriate value
    volumes:
      - /tmp/.X11-unix:/tmp/.X11-unix

Remember: environment needs DISPLAY=:0 in Windows and DISPLAY=host.docker.internal:0 in macOS. After editing and saving the file, open the browser along with the Playwright Inspector with this command:

docker compose run --rm web npx -y playwright open google.com

Most likely you already have a Playwright test suite that you run in CI or locally via Docker. Note that as long as you set the environment and volumes keys on your service definition you will be able to run Playwright in headed mode against your existing tests. Assuming you have a web service configured to run your tests AND have configured the environment and volumes keys as described above, you can drop into bash and run your tests graphically with:

docker compose run --rm web bash
# If using the JS package
npx -y playwright test --ui

# For other platforms (Python, etc.)
PWDEBUG=1 <command you use to run playwright>

This setup enables you to keep your dev and testing environment containerized while providing the flexibility of a graphical interface when needed.

Recent Articles

  1. A back hoe on the bank of the Suez, trying to free the Ever Given cargo ship
    Article post type

    Learn Grid Now, Container Queries Can Wait

    Take your time with new CSS, but don’t sleep on the essentials

    Several people have asked recently why container queries aren’t being used more broadly in production. But I think we underestimate the level of legacy browser support that most companies require to re-write a code-base.

    see all Article posts
  2. A clear kitchen blender filled with chopped fruit and greens
    Article post type

    Can you un-mix a mixin?

    Rethinking the CSS mixin proposal after CSS Day

    The CSS Working Group has agreed to move forward with CSS-native mixins. But some recent mixin-like CSS tricks have an advantage that the official proposal doesn’t account for: they make it easy to remove a mixin after it’s already been mixed in.

    see all Article posts
  3. Article post type

    Generating Frontend API Clients from OpenAPI

    API changes can be a headache in the frontend, but some initial setup can help you develop and adapt to API changes as they come. In this article, we look at one method of using OpenAPI to generate a typesafe and up-to-date frontend API client.

    see all Article posts