Learn about this first
Browser Use author Gregor Zunic publishes a practical guide describing six configuration methods for getting coding agents such as Claude Code, Codex, Gemini CLI, and Cursor to work in the user's real, already-logged-in Chrome. The core approaches are installing a browser extension to hand tabs over to the agent or enabling Chrome remote debugging so tools can connect to an open browser.
The guide says that when using Claude Code with a paid Claude plan, Claude in Chrome (claude --chrome) is the shortest path. However, the feature is unavailable when using an API key or claude setup-token, through Amazon Bedrock, Google Cloud Agent Platform, or Microsoft Foundry, and in WSL. Starting with Chrome 136, --remote-debugging-port and --remote-debugging-pipe are ignored when used with the default Chrome data directory; you need to use --user-data-dir to specify a different location.
The guide also warns that the command-line tool sends usage telemetry by default, including code passed to the agent and its output. This can be disabled with browser-harness telemetry disable.
AI-generated from reports · updated 22 days agoLatest developments
A hands-on guide to six ways to let coding agents use a logged-in Chrome
Event progress1
An introduction to methods for enabling coding agents to use logged-in Chrome
Browser Use: A practical guide to six ways of enabling coding agents to use an already-logged-in Chrome