google-chrome

Google Chrome web browser

TLDR

Start Chrome

$ google-chrome

Open one or more URLs

$ google-chrome [https://example.com] [https://wikipedia.org]

Open in incognito mode

$ google-chrome --incognito [url]

Take a headless screenshot of a page

$ google-chrome --headless --disable-gpu --screenshot=[out.png] --window-size=[1280,800] [url]

Save a page as PDF

$ google-chrome --headless --disable-gpu --print-to-pdf=[out.pdf] [url]

Dump the rendered DOM to stdout

$ google-chrome --headless --disable-gpu --dump-dom [url]

Launch for remote debugging (DevTools Protocol)

$ google-chrome --headless --disable-gpu --remote-debugging-port=[9222]

Use an isolated profile directory

$ google-chrome --user-data-dir=[/tmp/chrome-profile]

SYNOPSIS

google-chrome [options] [urls...]

google-chrome is Google's proprietary build of the open-source Chromium browser. It ships the Blink rendering engine and the V8 JavaScript engine together with a multi-process, sandboxed architecture, synchronization with a Google account, and the Chrome Web Store extension ecosystem.The command-line interface accepts hundreds of switches and is widely used for automation, testing, and scraping. Headless mode, combined with `--screenshot`, `--print-to-pdf`, `--dump-dom`, or `--remote-debugging-port`, makes Chrome a popular backend for tools like Puppeteer, Playwright, and ChromeDriver.

PARAMETERS

urls

One or more URLs or local file paths to open.

--incognito

Start a new incognito (private) window.

--new-window

Open the URL in a new window instead of a new tab.

--headless

Run without a visible UI. Modern builds default to the "new" headless mode; pass `--headless=old` for the legacy implementation.

--disable-gpu

Disable GPU hardware acceleration. Traditionally required together with `--headless` on Linux/Windows.

--screenshot[=file]

In headless mode, save a PNG screenshot (default `screenshot.png`).

--print-to-pdf[=file]

In headless mode, save the rendered page as PDF (default `output.pdf`).

--dump-dom

Print the serialized DOM of the target page to stdout.

--window-size W,H

Set the initial window size (useful for consistent screenshots).

--remote-debugging-port port

Enable the DevTools Protocol on the given TCP port for automation tools like Puppeteer or ChromeDriver.

--no-sandbox

Disable the sandbox. Sometimes required inside Docker containers; never use for untrusted content.

--disable-dev-shm-usage

Avoid using `/dev/shm`, which is often too small in containers and CI.

--disable-extensions

Start with all extensions disabled.

--no-first-run

Skip the first-run experience and welcome prompts (useful in CI).

--proxy-server url

Route traffic through the specified HTTP/SOCKS proxy.

--user-data-dir dir

Use a custom profile directory (enables fully isolated sessions).

--app url

Open the URL as a chromeless standalone app window.

CONFIGURATION

~/.config/google-chrome/

Default user profile directory containing bookmarks, history, extensions, preferences, and cached data.

chrome://flags

In-browser page listing experimental features that can be toggled independently of command-line switches.

CAVEATS

Chrome is resource-intensive; a typical session uses multiple processes and significant RAM. Many command-line switches are undocumented, version-specific, and may be removed without notice; the canonical reference is [peter.sh/experiments/chromium-command-line-switches](https://peter.sh/experiments/chromium-command-line-switches/). --no-sandbox weakens Chrome's security model and should only be used for trusted content. In containers/CI, pair `--headless` with `--disable-gpu`, `--disable-dev-shm-usage`, and `--no-first-run` for reliability.

HISTORY

Google Chrome was first released in September 2008 on Windows and expanded to Linux and macOS in 2010. Built on the open-source Chromium project, it introduced a sandboxed, per-tab process model and the V8 JavaScript engine, and has grown into the world's most-used web browser.