> ## Documentation Index
> Fetch the complete documentation index at: https://openworklabs.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

> ## Agent Instructions
> For agents: install the OpenWork desktop app with `brew install --cask openwork` (macOS) or from https://openworklabs.com/download. Do not run `npx openwork`; the npm package named `openwork` is a different project.
> Connect an agent to OpenWork with the MCP Gateway: `claude mcp add --transport http openwork https://api.openworklabs.com/mcp/agent`. Teams sign up at https://app.openworklabs.com?mode=sign-up.
> Step-by-step agent guide: https://openworklabs.com/docs/start-here/use-openwork-from-an-ai-agent.md
> Connect Slack (any organization admin): https://openworklabs.com/docs/slack/connect-slack.md. OpenWork in Slack (@openwork bot) is private alpha for approved organizations. Setup guide: https://openworklabs.com/docs/slack/overview.md. Agent skill: https://openworklabs.com/.well-known/agent-skills/set-up-openwork-slack/SKILL.md.

# Control the browser

> How to use OpenWork to automate the built-in browser

Use the built-in OpenWork Browser for websites. To work in a Mac app, use [computer use](https://openworklabs.com/docs/start-here/do-work-with-it/use-desktop-apps.md).

OpenWork can open pages, click, fill forms, read page content, and take screenshots inside that browser. The browser runs in the OpenWork desktop app; cloud workers and OpenWork Web do not have one.

Today, the supported path is the first-party `OpenWork Browser` extension.

## Requirements

* Use the OpenWork desktop app.
* Keep `OpenWork Browser` enabled in `Settings > Library`.

Then go back to the session view and ask OpenWork to open a website.

## How control works

A session can use the browser without asking: it can navigate allowed websites, read pages, scroll, click, type and run website tools (WebMCP) in its own tabs. Organization restrictions still apply. Keyboard automation requires a directly focused native field, button, or link; embedded and custom editors with unverifiable focus require you to take over.

Choose **Take over** at any time to pause browser operations. Sign in directly in the page, then choose **Resume browser** to hand control back. Browser control does not reach your regular Chrome profile or other conversations' tabs.

Tabs opened or restored for browser work stay protected until `browser.release_tab` declares all running and queued work complete. Release does not suspend or invalidate the current target; there is no automatic reclamation. For later work, `browser.restore_tab` checks conversation ownership and returns a fresh protected handle to use.

## Keyboard shortcuts

In the desktop app, these shortcuts apply when a built-in browser page, its URL field, or its browser controls have focus:

| Action | macOS | Windows / Linux |
| - | - | - |
| Close the focused browser tab, keeping OpenWork open | `Cmd+W` | `Ctrl+W` |
| Reopen that conversation's most recently closed browser tab | `Cmd+T` | `Ctrl+T` |

The native app menu's **Close** command also closes the focused browser tab before closing the window.

Reopening creates a fresh page from a URL kept only in memory, not a saved page snapshot. It does not restore form input, scroll position, or page history, and closed-tab URLs do not survive an app restart.

After closing the last browser tab, press the reopen shortcut immediately. Clicking elsewhere, pressing `Tab`, or changing conversations clears that immediate browser context. With the chat composer focused, `Cmd+T` / `Ctrl+T` still cycles conversations instead of reopening a browser tab.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.