Browser automation settings

Browser automation settings control the local browser engine that agents use when running browser tasks. Open them from Settings > Browser Automation in the desktop app. These settings apply to local agent runs — agents running on cloud desktops use the cloud desktop's own browser.

Enable browser

The master toggle to enable or disable the browser automation engine. When disabled, agents cannot perform browser tasks locally.

Headless mode

When Headless mode is on, the browser runs without a visible window. This is useful when you want agents to perform browser tasks silently in the background.

Turn it off if you want to watch the browser as the agent navigates.

Attach only

When Attach only is on, MultiClaw connects to an already-running Chrome session instead of launching a new one. This lets you run browser automation inside your existing browser profile.

Default profile

Choose whether agents use the OpenClaw-managed browser (an isolated profile) or attach to an existing Chrome session on your machine.

Browser executable path

If MultiClaw cannot find your browser automatically, enter the full path to the browser executable here. Leave blank to use auto-detect.

Advanced

The following settings are for advanced users and should only be changed if recommended by support:

• No sandbox — disables browser sandboxing. Use with caution.
• Browser accent color — sets a custom hex color for the browser UI chrome.
• Allow private network — controls whether the browser can access local network addresses.
• CDP timeout — timeout in milliseconds for Chrome DevTools Protocol connections.
• CDP handshake timeout — timeout in milliseconds for the CDP handshake.