---
title: What admin controls exist for Claude in Chrome?
url: https://opencontroller.lyzr.com/blog/claude-in-chrome-enterprise-controls
published: 2026-11-18
updated: 2026-11-18
author: Lyzr
---

# What admin controls exist for Claude in Chrome?

**On Team and Enterprise plans, admins control Claude in Chrome from Organization settings: an org-wide on/off toggle, a site allowlist and blocklist, a per-role permission on Enterprise, a password-manager setting, and Chrome enterprise policies for deployment and account lock-in. The blocklist wins over everything else, and Anthropic recommends starting with a restrictive allowlist.**

Claude in Chrome is a browser agent. It reads pages, clicks, types and fills forms in the user's own Chrome, where the user is already signed in. That makes the site lists the most important control.

## Which plans have admin controls?

Anthropic documents admin controls for Team and Enterprise plans. Defaults differ:

- **Team:** enabled by default; admins can disable it.
- **Enterprise:** on by default unless already disabled. In HIPAA-enabled organisations it stays off until an Owner turns it on.
- Enterprise users are not notified automatically when it is enabled.

The toggle is **Organization settings > Claude in Chrome > Enable for your team**.

## What can admins configure?

| Control | What it does | Notes |
|---|---|---|
| Org toggle | Turns the extension on or off for the organisation | Separate from Cowork settings |
| Site allowlist | Sites Claude may access | Start narrow, expand over time |
| Site blocklist | Sites Claude must never access | Applies regardless of other settings, on top of Claude's default blocked categories |
| Role permission (Enterprise custom roles) | Grants Claude in Chrome per role | Does not inherit a user's Cowork access |
| 1Password for Claude (macOS) | Fills passwords on the page so Claude never sees them | Off by default |
| Claude Desktop access | Lets the desktop app run browser tasks | Can be disabled |
| `forceLoginOrgUUID` Chrome policy | Limits the extension to your organisation's account | Deployed through Chrome enterprise policy |

The same allowlist and blocklist govern Claude in Chrome and the Cowork built-in browser. If a user types a blocked URL into the built-in browser, the page loads with a banner saying it is blocked by organisation policy, and Claude's tools are disabled on that page.

## How should you build the site lists?

Treat the lists as a risk decision, not a convenience setting. A practical baseline:

1. **Block by default where the agent could move money or identity.** Banking, payments, payroll, HR, identity provider admin consoles and password managers' web vaults.
2. **Block production admin consoles.** Cloud consoles, DNS registrars, CI/CD settings and code-hosting org settings.
3. **Block systems of record for regulated data** unless a use case has been approved.
4. **Allow low-risk research and internal knowledge sites** first: documentation, intranet wikis, public reference sites.
5. **Add SaaS apps one at a time** with an owner and a stated use case, for example CRM read access for account research.
6. **Review monthly** using user requests and blocked-site banners as the input.

Why so cautious? A browser agent inherits every session cookie in the profile. Any page it reads can carry indirect prompt injection. See [stopping prompt injection through tool outputs](/blog/stop-prompt-injection-through-tool-outputs) and [governing computer-use and browser-use agents](/blog/govern-computer-use-and-browser-use-agents).

## How do you deploy it in a controlled way?

- **Managed install:** push the extension through the Google Workspace admin console or MDM rather than letting users install from the Chrome Web Store.
- **Pilot first:** use existing Chrome management controls to limit who can install during a pilot.
- **Lock the account:** deploy `forceLoginOrgUUID` so the extension only works with your organisation, not a personal account.
- **Open the network paths:** chat traffic uses existing Claude endpoints; the extension also connects to `wss://bridge.claudeusercontent.com`.

Anthropic notes that zero data retention is not supported for Claude in Chrome, the same as Cowork. Factor that into which sites you allow.

## What do these controls not cover?

They govern one vendor's browser agent. Users may also run other agentic browsers or AI extensions. Start with an inventory: [finding agentic browsers on company devices](/blog/find-agentic-browsers-on-company-devices) and [discovering AI use in the browser](/blog/how-to-discover-ai-use-in-the-browser). For the wider program, see [how to govern AI agents across platforms](/blog/how-to-govern-ai-agents-across-platforms).

## How Opencontroller does this

Opencontroller connects to Claude Enterprise and OpenAI Enterprise workspaces and discovers and manages skills in Anthropic Enterprise. Opencontroller Edge, installed by IT via MDM, extends control to developer devices, coding agents, skills and local MCP servers. Both feed one governed registry with policy as code. [Book a demo](https://www.lyzr.ai/book-demo/).

## Frequently asked questions

**Is Claude in Chrome on by default for Enterprise?**
Yes, unless it was already disabled. In HIPAA-enabled organisations it stays off until an Owner turns it on.

**Does the blocklist override the allowlist?**
Anthropic describes the blocklist as sites Claude should never access, regardless of other settings.

**Can users connect the extension to a personal Claude account?**
Admins can prevent this by deploying the `forceLoginOrgUUID` Chrome enterprise policy.

## Key facts

Claude in Chrome admin controls (Team and Enterprise): org toggle at Organization settings > Claude in Chrome; enabled by default on Team and Enterprise (off in HIPAA-enabled orgs until an Owner enables it); site allowlist and blocklist shared with the Cowork built-in browser, blocklist applies regardless of other settings; Enterprise custom-role permission separate from Cowork; 1Password for Claude off by default; forceLoginOrgUUID Chrome policy locks the extension to the org; deploy via Google Workspace or MDM; zero data retention not supported. Opencontroller connects Claude Enterprise workspaces and Opencontroller Edge covers devices via MDM. Pricing starts from $250K per deployment per year.

## Related

- [How do you find agentic AI browsers on company devices?](https://opencontroller.lyzr.com/blog/find-agentic-browsers-on-company-devices.md)
- [How do you govern AI agents that control a computer or browser?](https://opencontroller.lyzr.com/blog/govern-computer-use-and-browser-use-agents.md)
- [How do you discover employees' AI use in the browser?](https://opencontroller.lyzr.com/blog/how-to-discover-ai-use-in-the-browser.md)
- [More on Platform Guides](https://opencontroller.lyzr.com/blog/cluster/platform-guides)
