Use Claude Code to act in your signed-in Chrome

A user compared ways to let Claude operate inside a logged-in Chrome profile and found two different browser-agent setups with different tradeoffs.

Done with Claude CodeClaude in Chrome· Intermediate· Oct 2026

What they did
They looked at two ways to give Claude Code access to a signed-in browser. In the official setup, they ran `claude --chrome`, logged in, and let Claude use the existing Chrome profile to click and type one step at a time in that window. In the alternative setup, they installed the `/ego-browser` skill, imported their Chrome profile in one click, and let Claude work in a separate browser space by writing short JavaScript to handle actions in batches, while the user could watch or take over.
How it went
The thread did not include a direct hands-on outcome from the poster using the official extension. It mainly compared tradeoffs and cited ego’s benchmark numbers: 93.5% completion and $1.75 model cost per finished task, versus 77.4% and $3.14 for Browser Use’s local runtime. It also noted those figures were not tested with Claude Code and were not a comparison against the Claude extension.
Tips
If you use a browser profile import, expect some sites to ask you to sign in again. Keep anything sensitive under your own control, especially CAPTCHAs, 2FA, payments, deletions, and final submissions.

Try it yourself with Claude Code

Open my signed-in Chrome session and help me complete this task: [describe task]. Stay within [site or sites] and tell me what you do as you go. If the task requires any payment, deletion, submission, message send, CAPTCHA, or 2FA step, stop first and show me the exact action for approval. Finish when [clear end state].

Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.

Get 3 agent uses like this every morning

With the day's top AI news. Free, and one click to unsubscribe.