Managed Chrome and Gemini
How to Disable Gemini in Chrome for Your Organisation
Gemini in Chrome reads the page your staff have open and, where its agent features are on, can click through sites inside their signed-in tabs. One policy switches it off, but only on the browsers and profiles that receive it.
Below: the policies, the values, the steps for each console, how to check the result and what stays open afterwards.
The values we suggest for a full stop. Some teams keep the assistant and switch off only the agent features; the table further down shows both options.
Next step
AI Agent Blocklist
Covers what the Gemini policy cannot reach: the Gemini web and desktop apps, agent extensions and other AI browsers, as ready files for Chrome policy, MDM, DNS filters and firewalls.
From $99 a month for one organisation
Updated every day. Download access within one business day of your order.
What Gemini in Chrome can reach on a managed device
Gemini in Chrome is not an extension you can find in the extension list. It ships inside Chrome, opens from a toolbar button or the side panel and works with whatever the user has in front of them. Its reach grows with every feature you leave switched on.
The open page
With the user's consent it uses the content of the current tab to answer. Intranet pages, CRM records, tickets and HR forms are pages too.
The side panel chat
Users type prompts next to the page. Prompts, page context and answers travel to Google's service to produce each reply.
Auto browse
Gemini works inside Chrome tabs to finish multi-step tasks the user hands over: opening sites, filling fields and moving through forms.
Saved sign-ins
With permission, auto browse can use Google Password Manager to sign in to sites during a task, so it acts with the user's own accounts.
Gemini Spark
Users can connect Chrome to Spark, Google's background agent, which can then run auto browse in the active browser session.
Shortcuts outside the toolbar
A system tray or menu bar entry and a keyboard shortcut can open Gemini even when the toolbar button is hidden.
Workspace users
Work accounts on qualifying Workspace editions follow your Admin console settings. If nobody changed them, Gemini integrations are allowed by default.
Personal profiles
Staff can add a second Chrome profile with a private Google account. Cloud settings for your domain do not follow that profile. Machine-level policies do.
New regions
Google decides availability per country, language and age. A launch in your country can switch the feature on for staff without any change on your side.
Signs that Gemini in Chrome is already in use
- A sparkle Gemini button at the top right of Chrome on company laptops
- chrome://settings/ai/gemini shows options the user can still change
- Activity in the Admin console reports for Gemini under Generative AI
- Helpdesk questions about auto browse prompts or connecting Spark
- Chrome profiles signed in with personal Gmail accounts on managed devices
- Staff pasting Gemini answers that quote internal pages they were viewing
Agent features change the risk from "what did the user share" to "what did the browser do". Our page on prompt injection in AI browsers explains how a web page can steer an agent that acts in signed-in tabs.
Gemini in Chrome removed? Why the button disappears
Staff often notice the change before IT announces it, and then search for why Gemini vanished. Five causes explain almost every missing Gemini button. The first one is the one you want to see on every managed device.
| Cause | What the user sees | Where to check |
|---|---|---|
| Policy set to Do not allow | No button, no side panel, Gemini settings no longer editable | chrome://policy lists GeminiSettings with value 1 |
| Unpinned by the user | Button gone, but the keyboard shortcut still opens Gemini | Right-click the top of the Chrome window and look for the pin option |
| Not eligible | The button never appeared | Country, Chrome language, age (18 or older) and whether the work edition qualifies |
| Signed out or Incognito | Missing in that window only | Gemini in Chrome needs a signed-in profile and is not offered in Incognito |
| Gemini app turned off for the account | Missing in the work profile, present in a personal one | The Gemini app service setting for that organisational unit or group in the Admin console |
A two-line note saves tickets: what was turned off, why, and which approved AI tool to use instead.
Admin console changes usually arrive within minutes, but Google allows up to 24 hours for every device.
If the button is gone in the work profile but back in a personal one, your policy is set at user level only.
The Chrome policies that switch Gemini in Chrome off
Google splits Gemini in Chrome into a main switch and separate switches for its agent features. We set all of them, so the intent stays clear even if someone later flips one back.
| Policy | What it controls | Values | To switch off |
|---|---|---|---|
| GeminiSettings | Gemini integrations in Chrome: the button, side panel and everything built on them | 0 Allow, 1 Do not allow | 1 |
| GeminiActOnWebSettings | Whether Gemini may act on web pages for the user (auto browse) | 0 Allowed, 1 Not allowed | 1 |
| GeminiActOnWebAllowedForURLs | Sites where Gemini may act even when the general rule says no | List of URL patterns | Leave empty |
| GeminiActOnWebBlockedForURLs | Sites where Gemini may never act | List of URL patterns | Banking, payroll and admin consoles if you allow acting at all |
| GeminiSparkSettings | Whether Gemini Spark may connect to Chrome and use auto browse | 0 Allowed, 1 Not allowed (the default) | 1 |
| GenAiDefaultSettings | Default for Chrome's covered generative AI features, Gemini in Chrome included | 0 Allow, 1 Allow without model improvement, 2 Do not allow | 2 (Admin console only) |
The agent policies have no effect while GeminiSettings blocks Gemini. Set them anyway: they keep the agent off if the assistant is ever switched back on.
A URL that matches both the allowed and the blocked list for acting is allowed. Keep the allowed list short and review it with each change.
If you want the chat but not the clicking, leave GeminiSettings at 0 and set GeminiActOnWebSettings and GeminiSparkSettings to 1.
AI Mode in the address bar, Help me write, tab organiser and Chrome's other AI features have policies of their own. We cover them on the page about Chrome AI policies in the Google Admin console.
How to disable Gemini in Chrome in each console
Use the console that already manages your browsers. If you run more than one, set the policy in each: a machine policy from Windows or macOS covers every profile on the device, a cloud policy covers managed users and enrolled browsers.
Google Admin console
- Sign in with an administrator account that can manage Chrome settings.
- Open Menu, then Devices, then Chrome, then Settings. User & browser settings opens. With Chrome Enterprise Core the path is Menu, Chrome browser, Settings.
- Select the organisational unit or group. Group settings override organisational units.
- Scroll to Generative AI and open Gemini integrations.
- Choose Do not allow Gemini integrations and click Save, or Override for a child unit.
- Repeat for the auto browse and Spark settings in the same section if they are listed for your edition.
Labels in the Admin console move from time to time. Check the current admin console if a heading has changed. For extension rules in the same console, see blocking AI extensions in the Google Admin console.
Group Policy or the registry
- Import the current Chrome ADMX templates. Older templates do not list the Gemini settings.
- In the Group Policy Management Editor, find Settings for Gemini integration under Google Chrome. Search the templates if the folder name differs.
- Enable it and pick Do not allow Gemini integrations.
- Do the same for the act on web and Spark settings.
- Without GPO, create a DWORD named GeminiSettings with value 1 under HKEY_LOCAL_MACHINE\SOFTWARE\Policies\Google\Chrome, plus GeminiActOnWebSettings and GeminiSparkSettings with value 1.
The same templates carry the extension policies; our guide to blocking Chrome extensions with Group Policy and the registry uses them.
Microsoft Intune on Windows
- Go to Devices, Manage devices, Configuration, then Create and New policy.
- Platform: Windows 10 and later. Profile type: Settings catalog.
- Click Add settings and search for Gemini. If your tenant does not list the Chrome Gemini settings yet, import current Chrome ADMX files or push the registry values with a script.
- Set Settings for Gemini integration to Do not allow Gemini integrations, add the agent settings, and assign the profile to device groups.
- Open the profile's per-setting status to find devices where it failed.
Intune menus change often, so check the current admin console. Extension rules in Intune are covered in block Chrome extensions with Intune.
macOS with Jamf Pro or another MDM
- Create a configuration profile with a custom settings payload for the Chrome preference domain com.google.Chrome. In Jamf Pro this is the Application & Custom Settings payload.
- Add GeminiSettings, GeminiActOnWebSettings and GeminiSparkSettings as integer keys with value 1.
- Scope the profile to every Mac that runs Chrome, shared and lab Macs included.
- On a test Mac, open chrome://policy and confirm the source shows Platform.
Payload names differ by MDM; check the current console. The same profile can carry extension rules, see block Chrome extensions on Macs with Jamf.
ChromeOS, iPhone, iPad and Android
- On managed Chromebooks, the Admin console setting above applies to managed users and devices. Gemini in Chrome needs a Chromebook Plus device in the first place.
- GeminiSettings is supported in Chrome on iOS, and in newer Chrome versions on Android. Push it as managed app configuration through your mobile MDM, or let the Admin console apply it to signed-in work accounts.
- On student devices, combine this with the controls in our guide to AI browsers on school Chromebooks.
Check that Gemini in Chrome is really off
A policy that never reached the device looks exactly like a policy that worked, until someone opens the side panel. Run these checks on one device per platform and per management method.
- chrome://policy: click Reload policies and find GeminiSettings with value 1 and status OK. Do the same for the act on web and Spark policies.
- Source column: Platform means GPO, registry or MDM. Cloud means the Admin console. A conflict between sources shows as a warning next to the policy.
- chrome://management: confirm that the browser or the profile is managed by your organisation.
- Gemini settings page: open chrome://settings/ai/gemini. The user should no longer be able to turn Gemini features back on.
- Second profile test: add a personal Chrome profile on the same test device. With a machine policy in place, Gemini stays off there too.
- After each Chrome update: new Gemini features often arrive with new policies. Recheck after major versions and read the Chrome Enterprise release notes for Gemini entries.
Gemini in Chrome in the EU, the UK, the US and for enterprise
Many admins search for the regional status first, hoping the feature is not offered where they are. Regional status helps you plan the message to staff. It should not decide whether you set the policy.
United States first
Google launched Gemini in Chrome for desktop users in the US in September 2025, then added countries and languages through 2026.
UK and EU
Reports in mid 2026 said UK desktop users had it while EU countries were still waiting. One announcement can change that.
Work accounts
Gemini in Chrome for work accounts needs a qualifying Workspace edition, and both the Gemini app setting and the Chrome Gemini integrations setting allowed.
Who can use it
Users must be 18 or older, signed in to Chrome and using a supported Chrome language. Gemini in Chrome is not offered in Incognito windows.
Platforms
Google lists Windows, macOS and Chromebook Plus computers, iPhone and iPad in supported regions, and Android in the US.
Data for Workspace users
Google states that prompts and tab context from Workspace users are not used to train its models and that Workspace data protections apply.
Geography is not a control. Staff travel, accounts move between regions and a regional launch arrives without a change ticket. Set the policy on every device now, including in countries where the feature is not offered yet.
Turning off Gemini in Chrome does not block Gemini
Google states that switching off the Chrome setting does not change the Gemini app on the web or on mobile. The policy closes one door in one browser. Staff who want an assistant or an agent have several others.
What the Chrome policy stops
- The Gemini button and side panel in managed Chrome
- Page content shared with Gemini in Chrome
- Auto browse inside Chrome tabs
- Gemini Spark driving the managed browser
What stays open afterwards
- The Gemini web app in any browser on the device
- The Gemini desktop app on Mac and Windows
- The Gemini command-line agent on developer machines
- Agent extensions such as Claude in Chrome and many smaller ones
- AI browsers installed next to Chrome, from our list of AI browsers
What we put in the files for this gap: block, monitor, allow
| Item | Example entry | Identifier type | Suggested action |
|---|---|---|---|
| Gemini web app | gemini.google.com | Domain | Allow or block by your Gemini app decision |
| Gemini desktop app, Mac | com.google.GeminiMacOS | macOS bundle ID | Block where Gemini is not approved |
| Gemini desktop app, Windows | Google.GeminiDesktop | Windows package ID | Block install through your MDM |
| Gemini CLI | @google/gemini-cli | Command-line package | Monitor on developer machines, block elsewhere |
| Agent extensions | Chrome and Edge extension IDs | Extension ID | Block with ExtensionInstallBlocklist |
| Approved AI tools | Your own choices | Any | Allow and record them in the AI use policy |
1
Pick the files for your tools: the Chrome policy file for extension blocking, the domain list for DNS or firewall, and the Mac and Windows identifiers for Jamf or Intune.
2
Filter by category or risk grade if you want only agents and AI browsers blocked while chat tools you approved stay open.
3
Import once, then refresh on a schedule. New agents, extensions and AI browsers are added to the list every day.
Keep Gemini in Chrome off, and the rest of the agent layer too
The Gemini policy costs nothing and is quick to set. The harder part is everything around it: new AI browsers, new agent extensions and new desktop agents that arrive every week.
The AI Agent Blocklist gives you that part as files you import into the tools you already run. Every entry is checked before it enters the list.
From $99 a month for one organisation
- 2,600+ AI agent domains: AI browsers, browser agents, autonomous agents, MCP servers, coding agents and more
- 5,300+ agent identifiers on devices: extension IDs, macOS bundle IDs, Windows package IDs, process names and CLI packages
- Tool name, category, identifier type and risk grade on every entry
- Plain domain list, CSV, JSON, hosts file, PAC file, firewall URL list and DNS filter imports
- Chrome and Edge extension block lists for Group Policy, Intune and the Google Admin console
Ask us or order a plan
Tell us what you need to block and where your rules live. We reply within one business day.
- Files for firewalls, DNS filters, secure web gateways, MDM and browser policies
- Updated every day as new AI browsers, agents and extensions appear
- Plans for one organisation, for MSPs and for product makers, or a hosted filter we run for you
Questions about disabling Gemini in Chrome
Can I disable Gemini in Chrome for everyone in my organisation?
Yes. Set the Chrome policy GeminiSettings to 1 (Do not allow Gemini integrations) in the Google Admin console, Group Policy, the registry, Intune or a macOS profile. Set it at machine level as well, so it also covers personal Chrome profiles on company devices.
How do I disable Gemini in Chrome with GPO?
Import current Chrome ADMX templates, open the setting Settings for Gemini integration under Google Chrome, enable it and pick Do not allow Gemini integrations. Without templates, create the DWORD GeminiSettings with value 1 under HKEY_LOCAL_MACHINE\SOFTWARE\Policies\Google\Chrome.
Was Gemini in Chrome removed, or is it hidden by my organisation?
Open chrome://policy. If GeminiSettings shows value 1, your organisation turned it off. If no policy is listed, the button may be unpinned, your account or region may not be eligible, or you are in Incognito or signed out of Chrome.
Does turning off Gemini in Chrome block the Gemini app?
No. Google states that turning off the Chrome setting does not change the Gemini app on the web or on mobile. The web app, the desktop apps, the command-line agent and AI extensions need their own blocks.
Is Gemini in Chrome available in the EU?
Google rolls it out country by country. Reports in mid 2026 said the UK had it on desktop while EU countries were still waiting. Check Google's current list and set the policy everywhere anyway, so a regional launch changes nothing on your devices.
What is the Chrome Gemini agent and can I turn it off separately?
Auto browse lets Gemini act in Chrome tabs to finish multi-step tasks, and Gemini Spark can connect to Chrome to use it. GeminiActOnWebSettings and GeminiSparkSettings set to 1 turn these off while the assistant itself stays available.
Does Gemini in Chrome work on ChromeOS and mobile?
Google lists Chromebook Plus devices, iPhone and iPad in supported regions, and Android in the US. GeminiSettings is supported on ChromeOS and iOS, and newer Chrome versions on Android, so the same off switch applies there.
Google, Chrome, ChromeOS, Chromebook, Android, Gemini and Google Workspace are trademarks of Google LLC. Claude is a trademark of Anthropic, PBC.
Microsoft, Windows and Intune are trademarks of Microsoft Corporation. Jamf and Jamf Pro are trademarks of JAMF Software, LLC. Apple, Mac, macOS, iPhone and iPad are trademarks of Apple Inc.
Names are used only to identify the products discussed. AI Agent Blocklist is not affiliated with, sponsored or endorsed by these companies.