# Manus AI Browser Operator: Setup, Logins, Limits and Privacy

> How Manus AI's Browser Operator works in your own Chrome or Edge: setup, Take Over and saved logins, phone and API use, limits, and what reaches Manus.

_Source: https://professionalstoolkit.com/articles/manus-ai-browser-operator — The Professional's Toolkit · updated November 2026_

---

Manus AI's Browser Operator is a Chrome or Edge extension that lets the agent work inside your own desktop browser, with the sites you are already signed in to and your own IP address, instead of the cloud browser that runs inside each task's virtual machine. It has been open to all users since November 2025, and you switch it on through the My Browser connector. **The short answer:** use it for logged-in, security-sensitive or bot-wary sites, and leave public pages to the cloud browser. The trade is real. What Manus reads in those tabs goes to its servers.


**Per session** — Each Browser Operator session needs your authorization, and closing the agent's tab stops the task



- **Chrome, Edge** — Browsers Manus recommends for the extension today
- **Per session** — Each Browser Operator session needs its own authorization
- **Nov 2025** — Browser Operator opened to all Manus users


## Cloud browser or My Browser: which one does the work

Every Manus task gets its own cloud machine, and by default the agent browses from there. That browser runs on data center IP addresses, so some sites answer with CAPTCHAs or flag the visit as automated [1]. When a CAPTCHA, an SMS code or a multi-factor prompt appears, Manus asks you to Take Over, you clear the hurdle yourself, and control goes back to the agent. Those cloud sessions are encrypted and isolated per user. For public research this is usually enough, and it keeps your own computer out of the loop entirely.

My Browser flips the arrangement. The extension drives tabs on your desktop, using the cookies, sessions and local IP you already have, so a site sees an ordinary visit from your machine rather than a login from a data center. Manus recommends it over the cloud browser for banking and other security-sensitive sites. The price of that convenience is twofold. Your computer has to stay on and online for the whole task, and the agent sees whatever those logged-in pages show, including paid services you subscribe to.

| | Cloud browser | My Browser (Browser Operator) |
|---|---|---|
| Where it runs | The task's cloud virtual machine | Your desktop Chrome or Edge |
| IP address | Data center addresses | Your local address |
| Getting past a login | Take Over, then optionally save the login | Your existing sessions |
| Passwords | Saved logins kept only if you agree | Manus does not see or store them |
| Your computer must be on | No | Yes |



> ⚠ **Authorized pages reach Manus servers:** The Privacy Policy says page content from authorized sites, including logged-in paid services, plus clicks, form filling and your existing login sessions, is sent to Manus servers. Only your passwords are not seen or stored.

## Turning it on: a toggle, an extension and one click per session

Setup lives in the Connectors tab, the same place you connect Gmail or GitHub, and My Browser sits there among the built-in app connectors next to Google Workspace, Outlook and Shopify. Like any connector, Manus may suggest it during a task but cannot enable it without your approval, and it can only do what your signed-in accounts already allow. The sequence Manus documents has four steps:

1. Open Connectors in Manus and toggle on My Browser.
2. Install the Browser Operator extension in Chrome or Edge.
3. Give Manus a task that needs a site you are signed in to.
4. Click Authorize when Manus asks to take over a new tab.

Every session needs its own authorization, so yesterday's approval does not give the agent standing access to your browser today. The agent works in a dedicated tab group you can watch, all of its actions are logged, and closing that tab stops the task [2]. Chrome and Edge are the recommended browsers right now. Manus says support for others is coming, which means anyone on a different browser should not plan a workflow around it yet. Keep the extension installed only on the machine you actually want the agent to use.


> 💡 **Route sensitive sites through My Browser:** Manus recommends My Browser over the cloud browser for banking and security-sensitive sites, because it uses your local IP and existing sessions. Page content from those sites is still sent to Manus servers; only your passwords are not seen or stored.


## Logins, CAPTCHAs and what happens to your passwords

The two browsers treat credentials differently, and it matters which one a task is using. In the cloud browser, Take Over hands you control of Manus's own computer, so you can type a password, solve a CAPTCHA or enter a code, then pass the keyboard back. After a manual sign-in, Manus asks whether to save that login for future tasks. You can switch that off in Settings under Cloud Browser and manage saved logins site by site. The Privacy Policy says saved third-party credentials are used only to authenticate and to carry out the actions you request.

Browser Operator never needs that step. Because it reuses sessions already open in your browser, Manus says it does not see or store your passwords at all. Do not merge the two promises in your head. A login saved in the cloud browser is stored by Manus. Through My Browser, Manus does not see or store your passwords, but it uses your existing sessions and sends what it reads on those pages to its servers. Removing the My Browser connector clears its login information and saved credentials, which is the cleanest way to reset if a session ever looks wrong. It does not delete what Manus already holds in conversations, task files or memory. **Pick the browser by which sites need your own sessions**, not by habit.

## Starting browser tasks from a phone, another computer or the API

The extension always runs on a desktop browser that must be online. From the Manus mobile app you can start and monitor a task, and a prompt sent from your phone can open a browser session on your authorized desktop that runs in the background. Nothing runs in the phone's own browser. If you work across several machines, the Allow Cross-Browser Tasks toggle lets Manus drive any authorized browser session linked to your account, so a task can begin on one device and continue on another without starting over.

Preferred Browser narrows that choice. It sets one authorized Chrome browser as the default place where the agent works, which requires the extension on that computer, with Chrome left open and connected. You can then start tasks from another computer signed in to the same account. Manus notes that this setting may not appear on every account yet.

Developers can reach the same flow over the API: a waiting event asks for a browser, a list call shows which clients are online, and a confirmation picks one. An empty list means the extension is missing or switched off. Manus pitches this for internal tools that have no public API of their own. If you want the agent to work across local files and apps rather than only browser tabs, our [Manus Studio desktop app guide](/articles/manus-ai-desktop-app) covers Computer Use and its permission modes.

## Where it stumbles, and when to step in

Manus is candid about the rough spots. Drag-and-drop and multi-step forms may not work perfectly, and sites with aggressive anti-bot measures can still need you to take over by hand, even from your own IP. The agent pauses at sensitive steps such as a payment screen and waits for your confirmation, and you can interrupt at any moment. That makes Browser Operator a poor fit for unattended checkouts and a good fit for read-heavy work across accounts you already use, such as pulling figures from dashboards that sit behind a login.

That read-heavy pattern scales. Manus describes pairing Browser Operator with Wide Research to read every product page and review in a list and return a comparison table. Wide Research itself is for paid users only and starts automatically when the agent decides a job can be split, as our [guide for researchers and analysts](/articles/manus-ai-for-researchers) explains. Browser tasks spend credits like any other task, and Manus cannot tell you a task's credit cost before it runs. Keep each browser job tightly scoped. A specific request with one site and one output costs less than a vague errand.


> ⚠ **Forms and anti-bot sites can stall:** Drag-and-drop and multi-step forms may not work perfectly, and sites with aggressive anti-bot measures can still need you to take over by hand.


## What you hand over when you click Authorize

The Privacy Policy is plain about the data flow. Turning on the extension and the My Browser connector gives Manus the page content of authorized sites, including logged-in premium services, plus clicking, form filling and your existing login sessions, and that data is sent to Manus servers [3]. You can revoke access by disabling the connector or uninstalling the extension, and local browser access stops immediately. Revocation is a brake for the future. It does not reach back.

The Terms set the scope. You authorize Manus to act on connected browsers and accounts within the permissions you choose, third parties may attribute the agent's actions to you even when you did not foresee them, and revoking a permission does not undo an action already completed [4]. If you instruct or approve a purchase, the agent may submit the order and payment for you. On a Team account, the Owner and Super Admins can access members' tasks. Our [buyer's checklist for Manus AI](/articles/manus-ai-buyers-checklist) puts these points beside credits, refunds and data deletion.


> 💡 **Revoke access when the job ends:** Disabling the My Browser connector or uninstalling the extension stops local browser access immediately. Actions already completed are not undone, and content Manus already holds in conversations, task files or memory is not deleted.



## FAQ

**Is Browser Operator available on the Manus AI Free plan?**

Manus says Browser Operator is available to all users, Free included. On Free, agent tasks run in Manus 2.0 Lite.

**Can collaborators in a shared Manus task use my browser?**

No. Manus disables connectors automatically once collaboration is enabled on a task, and My Browser is a connector.

**What happens if my Preferred Browser computer is offline?**

Manus can fall back to another available authorized browser on your account. Preferred Browser itself may not be available on every account yet.

**Can a Manus project use My Browser for every task in it?**

Connectors in Projects is available to all users and supports My Browser, so you can attach it at the project level. Each browser session still needs your authorization.


## The bottom line

Browser Operator is the right switch when a task needs your real sessions: sites behind a login, dashboards you pay for, or pages that block data center traffic. It is open to every Manus user, takes four steps and one click per session, and stops when you close its tab. Its limits are practical ones, from shaky drag-and-drop to an always-on desktop. The serious question is trust. Every authorized page reaches Manus servers, and every action can be attributed to you. Turn it on per job, watch the tab group, and revoke it when you are done.



## Sources

[1] Manus docs: Cloud browser — https://manus.im/docs/features/cloud-browser (2026-11)
[2] Manus docs: Browser Operator — https://manus.im/docs/features/browser-operator (2026-11)
[3] Privacy Policy — https://manus.im/privacy (2026-11)
[4] Terms of Use — https://manus.im/terms (2026-11)
[5] Manus plans and pricing (US) — https://manus.im/pricing
[6] Manus Help Center — https://help.manus.im/en/
[7] Manus documentation — https://manus.im/docs/introduction/welcome