How to use browser automation

Support Help
1

I want the chat to be able to see the browser, the HTML code, and the console, and to be able to navigate through it — basically what is described in the Cursor documentation. But no matter what I do, the chat can’t see my browser, regardless of whether it’s the built-in browser tab or the bundled Chrome.

image
image857×520 12.7 KB

image
image348×623 11.6 KB

Browser tab and bundled chrome work corretly, but agent cant reach it.

How to do it in correct way step by step?

2

Hey, thanks for the report.

This is a known issue, it’s a regression in versions 2.3.18+. The team is working on a fix. Browser automation runs through built-in tools and shouldn’t show up as a separate MCP server in Settings.

To help the team speed up the fix, can you share:

  • Help > Toggle Developer Tools > Console, any errors when trying to use the browser tools
  • The Request ID from the failed call (chat menu > Copy Request ID)

Related report: Browser MCP Tools Not Available

3

Hey, could you turn off Privacy Mode and copy the request ID? The team is trying to look into this. Thanks.

4

request ids: (from oldest to newest) (i dont have this one from screen)
ea83780c-416a-4709-838a-7da60be24879
c36efad0-8de3-41fc-8866-739a98767472
88e61407-90f2-4170-bb44-ff4b4db8045f

in .txt file logs from console.

logs.txt (54,7 KB)

1 Like