Use your own browser
Let the agent work in your own Chrome, Edge or Brave, where you're already signed in, through the Multilo extension. Chat with Multilo beside any page.
The built-in browser keeps its own sign-ins. When a task needs a site you’re signed in to in your usual browser, such as a university library, a journal’s submission system, Google Docs or a reference manager on the web, connect that browser instead. The agent opens its own tabs there, in a tab group named Multilo, and works in them while you watch.
Before you begin
- The Multilo app is installed, open and signed in. The extension works through the app, so the app must be running.
- Your browser is Google Chrome, Microsoft Edge, Brave, or another Chromium browser such as Arc, Vivaldi or Opera.
- The extension and the app use the same Multilo account.
Set up
Open the setup
In Multilo, open Settings › Agent › Your browser and click Set up. The Use your browser with Multilo window lists the steps and ticks each one off as it’s done.
Add the extension
Click Open the Chrome Web Store and add the Multilo extension. In Edge, first allow extensions from other stores when Edge asks.
Sign in to the extension
In your browser, click Multilo in the toolbar to open its side panel, then click Sign in. Use the same account as the app.
Finish
When every step is ticked, click Continue with browser.
If the agent asks to use your browser before you’ve set it up, a card says “Multilo wants to use your own browser, where you’re signed in. Set up the Multilo extension for it?” Click Set up to open the same steps, or No, and don’t ask again. When you finish setting up from a chat, the agent carries on in your browser.
Settings › Agent › Your browser shows each browser and its state, such as “Connected as …”. If two browsers are connected, choose one in Browser to use.
Ask the agent to use your browser
Name the browser in your request, for example:
In Chrome, open my library account and list the books due this week.Use my browser to download the PDFs of the papers on my reading list from the journal site.The agent can open pages, read them, click, type, fill in forms, upload files, take screenshots and record a GIF of what it did. It uses the same tools as in the built-in browser. See Agent tools.
While it works in a tab, the tab has a purple edge and a bar that says Multilo is working in this tab. Click Stop in that bar to stop the chat.
Which tabs the agent uses
- The agent works only in tabs it opened, in the Multilo tab group. It can’t see your other tabs.
- To let it use a tab you opened, drag the tab into the Multilo group. Every chat can then use it.
- Deleting the chat that opened a tab closes the tab too.
Chat from the side panel
Click Multilo in the browser’s toolbar, or press ⌥⇧M (Windows: Alt+Shift+M), to open the side panel.
- Type in Ask Multilo… to ask about the page you’re on. This page is attached; click its × to leave it out.
- Answers appear as they’re written. Each side-panel chat is an ordinary chat in the app: Recent chats lists them, and Open this chat in Multilo opens it there.
- Select text on a page, right-click it and choose Ask Multilo about this to start a chat about the selection.
Sites, sign-ins and safety
- Each website asks first. The first time the agent uses a site in a chat, a card asks. Choose Allow for this chat, or Always allow for that site. This works the same as in the built-in browser.
- Sign in yourself. The agent uses the sites you’re already signed in to. It never types passwords or payment details. When a page asks for them, sign in, then tell the agent to carry on.
- Treat pages as untrusted. Text on a web page can try to give the agent instructions. The agent treats what it reads as information, and you approve each new site. Review what it does on sites where it can change things.
- Your account only. The app uses a browser only when the extension is signed in to the same Multilo account as the app.
Uploading files
When you ask the agent to attach a file to a page:
- An upload can be at most 10 MB in all.
- Files that hold credentials are never uploaded, such as
.env,.pemand.keyfiles, or anything in.ssh. Upload those yourself if you really mean to. - A file with several hard links isn’t uploaded. Copy it and ask again.
- Your browser must allow the extension to read files. Open
chrome://extensions, click Details under Multilo, and turn on Allow access to file URLs.
Turn it off
- For yourself: turn off the switch in Settings › Agent › Your browser. The agent then won’t use your browser until you turn it on again.
- For an organisation: set the policy
yourBrowser.enabledtofalse(on Windows, Allow the agent to use the user’s own browser in Group Policy). The feature is then off and people can’t turn it on. See Managed settings.
Troubleshoot
| You see | What to do |
|---|---|
| The side panel says to open the Multilo app | Open the app, then click Reconnect in the side panel |
| Settings › Agent › Your browser says “Not connected.” | Check that the browser is open with the extension turned on and signed in, then click Reconnect |
| “Signed in as …, not this app’s account” | In the side panel, click Sign in with another account and use the app’s account |
| “Sign in from the extension’s side panel” | Open the side panel and click Sign in |
| The agent stops responding in a tab | Look for a pop-up on the page, such as an alert, and close it, then tell the agent to carry on |
| An upload is refused because the browser needs file access | Turn on Allow access to file URLs for Multilo in chrome://extensions |
| You use more than one browser profile | Use the profile where the extension is installed and signed in |