Headed chrome in docker for automation and privacy

Launch headed Chrome in a Docker container running VNC, and get the remote debugging websocket endpoint back. This program solves the problem of not being able to run Playwright in headless=false mode from a Docker container. Additionally, these browser instances are completely ephemeral and therefore ideal for private browsing.

By using this, you agree to Google Chrome terms: https://www.google.com/chrome/terms/

Install and build the Docker image

Simply clone and build the docker iage.

git clone https://github.com/gbiz123/docker-chrome-vnc/
cd docker-chrome-vnc
docker build -t chrome-gui .
chmod +x run-chrome-gui.sh

Launch

Run the script to get the container ID and the websocket URL. The websocket URL can be used for Playwright automation over Chrome CDP.

./run-chrome-gui.sh 9222 5900 # Specify port for remote debugging
>>> {"containerId":de66831e80a5c0c6645e8240276549ea40ca18a30e816a100c5c28c4e139bfe3,"wsEndpoint":ws://localhost:9222/devtools/browser/8c6ff95f-f609-49eb-b701-d46b3e0e1c89}

Connect VNC

Use any VNC client to visualize and interact with the browser.

vncviewer localhost:5900

TO-DO

• [] Parameterize Xvfb display options to produce unique, yet reproducible canvas fingerprints
• [] Convert the run-chrome-gui.sh script to .bat file for Windows compatibility