Missing “Chrome” Option in Connect to Browser (Tools & MCP Settings)

Where does the bug appear (feature/product)?

Cursor IDE

Describe the Bug

The “Chrome” option is missing from the Connect to Browser setting under Tools and MCP. Previously, I was able to connect Cursor to Chrome for Browser Tools. Now the only available options are:

  • Off
  • Browser Tab

The “Chrome” option does not appear at all.

Steps to Reproduce

  1. Open Cursor IDE
  2. Go to Settings
  3. Navigate to Tools and MCP
  4. Locate Connect to Browser dropdown
  5. Observe that only “Off” and “Browser Tab” are available

Expected Behavior

The dropdown should include the Chrome option (as it did previously), allowing direct connection to the Chrome browser for Browser Tools.

Operating System

MacOS

Version Information

Version: 2.4.37
VSCode Version: 1.105.1
Commit: 7b9c34466f5c119e93c3e654bb80fe9306b6cc70
Date: 2026-02-12T23:15:35.107Z
Build Type: Stable
Release Track: Default
Electron: 39.2.7
Chromium: 142.0.7444.235
Node.js: 22.21.1
V8: 14.2.231.21-electron.0
OS: Darwin arm64 24.6.0

For AI issues: which model did you use?

Not model-related (Browser Tools issue)

For AI issues: add Request ID with privacy disabled

Not model-related (Browser Tools issue)

Additional Information

  • This previously worked in my setup.
  • No configuration changes were intentionally made.
  • Chrome is installed and functioning normally on the system.
  • Unsure if this was removed in a recent update or if it is a regression.

Does this stop you from using Cursor

Sometimes - I can sometimes use Cursor

Hey, this is actually expected behavior. The “Chrome” option was replaced with “Browser Tab”, which is Cursor’s built-in native browser. It works out of the box with no extra setup and is the recommended way to use the browser tools.

Docs: Browser | Cursor Docs

If you need to control an external Chrome or Chromium instance (for example, to test with extensions, authenticated sessions, or special browser profiles), you can set up the Playwright MCP server. Add this to your MCP config .cursor/mcp.json:

{
  "mcpServers": {
    "playwright": {
      "command": "npx",
      "args": ["-y", "@playwright/mcp@latest"]
    }
  }
}

This gives the Agent full browser automation features (navigation, clicks, typing, screenshots, etc.) through an external Chromium instance. More on MCP setup: Model Context Protocol (MCP) | Cursor Docs

Let me know if you want help setting up either option.

Thanks for your quick reply. Does the method you mentioned support Javascript execution as a tool?

The built-in browser tab doesn’t have a direct “run JavaScript” tool. It only supports navigation, clicking, typing, scrolling, screenshots, console output, and network traffic monitoring. That said, the Agent can read console output and logs, so you can ask it to watch for JS errors and debug messages.

If you need to run arbitrary JavaScript on the page, use the Playwright MCP server. It provides the browser_evaluate tool, which lets the Agent run JS directly in the page context, for example, to read DOM values, call page APIs, or change state.

My agent is not even able to make actions like clicking, navigating or anything, it says it doesn’t have tools related to that. @deanrie
Note that I am the latest version and cursor is able to connect to the native browser tab.

Sounds good. I am trying to setup the Playwright MCP.

I added the json config and I try to use it but it never seems to spin up the browser. Its stuck on “waiting for approval” but I do not see any place to “Approve”.

Thanks for the info.

Two things worth checking:

  1. Cursor Settings > Agents > make sure “MCP Tools Protection” is turned off. When it’s on, every MCP call needs manual approval, and if the approval UI doesn’t render, you’ll get stuck.

  2. If that didn’t help, try switching Auto-Run Mode to “Run Everything” in the same place. This fully skips manual approval.

Related threads:

Let me know if that helps.

1 Like

“Run everything” worked! Thanks for your help!

1 Like