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

# Browser Control

> Read and operate real browser tabs through the Enconvo Companion extension.

Browser Control lets an Agent work with browser tabs that have the Enconvo Companion extension connected. It can read a page, click controls, fill fields, inspect results, and manage tabs using the browser's existing sign-in state.

## Set up

1. Install [Enconvo Companion from the Chrome Web Store](https://chromewebstore.google.com/detail/enconvo-companion/afaklbihknigbnjdkegbnenkheogejkd) in a supported Chromium browser, or use the appropriate extension package for your browser.
2. Keep Enconvo and the browser running.
3. Open the extension popup and confirm its **Connected** status.
4. Check the Agent's browser tool access and any website approval requested by the extension.
5. Start with a read-only task, such as “Read this page and summarize the main points.”

See [Browser Extension](/features/browser-extension) for connected-browser status, targeting, and approval behavior. Safari does not use the Chromium companion extension; use the supported route for your browser rather than assuming every browser has identical capabilities.

## Manual Chromium installation

If your build provides the unpacked extension, open `chrome://extensions` (or your browser's equivalent), enable Developer mode, choose **Load unpacked**, and select `~/.enconvo/chrome_extension/`. Restart Enconvo if its bundled extension folder has not yet been prepared.

## Work with a specific tab

Name the browser and intended page in your request. For example:

> In Edge, read the current documentation tab and list the three setup steps. Keep the page open.

The Agent inspects page state, chooses an element, acts, and checks the resulting state. Page references can change after navigation or dynamic updates; a fresh snapshot helps recover when an element is no longer available.

## Available operations

| Area | Examples |
| - | - |
| Navigation | Open a URL, go back or forward, reload, wait for a condition |
| Reading | Page state, visible text, attributes, screenshots, selected content |
| Interaction | Click, type, fill, select, press keys, scroll, drag |
| Browser organization | Inspect and manage tabs, windows, and connected sessions |
| Advanced tools | Page evaluation, cookies, storage, history, and browser-specific inspection |

Capabilities and website access vary by browser, page, extension version, and enabled tools. Browser-internal pages and restricted content may not be accessible. Browser control can act on your real account; specify whether a task should only prepare a draft or also submit it.

## Configuration and diagnostics

Choose the default browser in Computer Use preferences, or name a browser in the task. The `browser-use/status` Local API route reports connection state. Developers can inspect the installed route schemas under **Settings → Developer → APIs** rather than relying on a fixed list of action names.

## Troubleshooting

* **Not connected:** verify that the extension is enabled and Enconvo is running; inspect the popup status. The companion connects locally on port 11225.
* **Wrong browser or tab:** explicitly name the target and request a fresh page state.
* **Click or fill fails:** inspect the updated page; a dialog, navigation, or embedded frame may have changed the target.
* **Blank screenshot:** make the target tab visible and wait for page content to load.
* **Website asks to sign in:** complete the site's normal login flow; an existing browser profile does not guarantee a valid session for every site.

Use [Computer Use](/features/computer-use) for native macOS apps, file dialogs, and desktop controls.


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