How to block AI music and voice generators on your network
AI tools can now write a full song from a text prompt or clone a person's voice from a short recording. Both are impressive, and both raise questions that schools and businesses need to answer before the tools appear on their networks.
This guide shows network admins how to block or limit these tools with domain rules, AI category rules, firewalls, MDM and browser policy.
AI music generators
Turn a prompt or lyrics into a finished track with vocals. The main concerns are copyright, coursework integrity and time on task.
AI voice generators
Text-to-speech, voice changers and voice cloning. The main concerns are impersonation, consent and personal data.
Why organisations restrict AI audio tools
Very few admins want to ban every AI feature. The goal is usually to keep high-risk tools away from people who do not need them, and to make approved tools the default.
Voice impersonation
A cloned voice can be used to imitate a manager, a teacher or a parent on a call. Businesses restrict cloning tools to reduce the risk of payment and access fraud.
Consent and personal data
Cloning needs a recording of someone's voice. Uploading a colleague's or student's voice without consent raises data protection questions in many countries.
Bullying and harassment in schools
Students can make a fake recording of a classmate or teacher saying something they never said. Schools block cloning tools on student networks for this reason.
Copyright and licensing
AI music services have faced copyright lawsuits from record labels. Companies that publish audio may want only cleared, approved tools on their networks.
Coursework integrity
Music and media courses need to know what a student made and what a tool made. Many teachers want these tools off unless a task allows them.
Uploads of company data
Scripts, unreleased product names and recordings pasted into a free tool leave your control. Approved tools with business terms reduce that risk.
Accessibility comes first. Text-to-speech helps many learners and employees. Keep approved speech tools and built-in device readers available while you restrict the rest.
Three families of AI audio tools
Grouping the tools this way makes policy easier, because each family has a different risk and a different set of people who may need it.
Song and music generators
Generate songs, beats or backing tracks from a prompt, sometimes with AI vocals.
Usually blocked for students; sometimes allowed for music staff.Text-to-speech and narration
Turn written text into natural-sounding speech for videos, training or accessibility.
Often allowed through one approved service.Voice cloning and changers
Copy a specific person's voice, or change a voice in real time on calls and streams.
Highest risk; restricted for almost everyone.Domains of well-known AI audio tools
These are the official domains of a few of the best-known services. Block each with all subdomains. This field moves quickly, so names and hosts change and new tools launch every week.
| Domain | What it is | Action |
|---|---|---|
suno.com | AI song generator. The service has also used an earlier .ai address; block both if you see them in logs. | Block |
udio.com | AI music generator that creates tracks from text prompts. | Block |
elevenlabs.io | AI voice platform covering text-to-speech, voice cloning and AI music. | Block or approve |
| Smaller AI audio tools and wrappers | Hundreds of sites that offer songs, covers, voices or voice changers, often built on larger services. | Category rule |
| Shared cloud and CDN hosts | Large hosting providers serve many unrelated services from the same domains. | Do not block |
| Your approved speech service | The tool your organisation has chosen for accessibility or narration. | Allow list |
Check any domain from your logs in the domain category lookup before you add it. For a wider view of the AI landscape, see the AI / LLM Tools category and the AI tools domains list.
A policy matrix: who gets what
Write the decision down by group before you touch any settings. This example works for many schools and mid-sized businesses. Adjust it to your own policy, then mirror each column as a filtering group so the rules match the paper.
Music generators: blocked.
Voice cloning: blocked.
Text-to-speech: approved tool only.
Music generators: on request for music and media staff.
Voice cloning: blocked.
Text-to-speech: approved tool.
Music: approved licensed tool.
Voice: approved narration tool.
Cloning: only with written consent.
All AI audio tools blocked by category.
Approved speech tool allowed.
Exceptions by request.
Questions to ask before approving a tool
An approved tool is only safer than the rest if someone checked it. These questions help IT, data protection and teaching or business leads agree quickly.
Terms of use
Who owns the output, and may it be used commercially? Are there limits on use by minors?
Data handling
Are uploaded recordings and scripts kept, and for how long? Can the organisation delete them?
Consent features
Does the tool require proof of consent before a voice is cloned? Can cloning be switched off for your accounts?
Admin controls
Is there a business or education plan with central accounts, so access can be removed when someone leaves?
Labelling
Can generated audio be marked as AI-made, so listeners know what they are hearing?
Review date
When will the approval be checked again? Terms and features in this market change often.
Once a tool passes, add its domains to the allow list for the right groups only, and record who approved it and when.
How to block AI music and voice tools
Combine a category rule for breadth with domain rules for the best-known names. Then add device controls where you manage devices.
DNS and category filter
Every device on your network- Hand out your filtering resolver through DHCP on every SSID and VLAN.
- Block the AI tools category for groups that should not use them, with allow-list entries for approved services.
- Add suno.com, udio.com and elevenlabs.io as wildcard blocks where your policy requires.
- Redirect outbound port 53 to your resolver and block DNS over TLS on port 853.
- No resolver of your own? A managed DNS filter covers this layer.
Firewall
Second check and app control- Mirror the AI category and domain rules in a URL filter profile.
- Use application control signatures for AI tools where your firewall offers them.
- Block QUIC on UDP 443 for filtered segments so server names are visible.
- Apply TLS inspection only where policy allows it, with sensitive categories excluded.
Browser policy
Managed laptops and Chromebooks- In Google Admin, add the domains to the URL block list for the relevant organisational units.
- In Microsoft Intune, push an Edge URL block list or a web content filter for the AI category.
- Restrict extensions to an approved list; many voice changers ship as extensions.
- Turn off browser DNS over HTTPS by policy.
MDM
School and company phones- Use an approved app list so AI song and voice apps cannot be installed.
- Remove existing copies from supervised iOS and Android Enterprise devices.
- Push a DNS or content filter profile so rules apply on mobile data.
- Prevent VPN profiles and private DNS changes.
App admin settings
AI inside approved software- Check your video, office and design apps for built-in AI voice or music features.
- Turn them off, or limit them to approved groups, in each product's admin console.
- Record the decision alongside your network policy so both stay in line.
For a wider AI plan, see how to block AI chatbots and how to block AI image generators.
Websites, apps and built-in features
AI audio reaches users through three routes. A network rule covers the first two well on your Wi-Fi; the third needs admin settings.
| Route | On your network | Off your network | Best control |
|---|---|---|---|
| Website in a browser | Blocked by domain and category rules | Blocked on managed browsers only | Category rule plus browser policy |
| Mobile or desktop app | Blocked when its domains are blocked | Not blocked unless the device is managed | MDM app allow list |
| Browser extension | Blocked if it calls blocked domains | Blocked if extensions are restricted | Extension allow list |
| Feature inside an approved app | Often allowed, because it uses the app's own domains | Same | The app's admin console |
Handling bypass attempts
Popular AI tools attract workarounds as soon as they are blocked. Plan for these before you switch the rule on.
Wrapper and mirror sites
Many small sites resell access to bigger AI audio services under new names. They appear weekly; a daily-refreshed AI category is the practical defence.
VPN apps and proxy sites
Block the VPN / Proxy / Anonymizers category and VPN signatures, and stop VPN profiles on managed devices. See how to block proxy sites.
Encrypted DNS in browsers
Turn off browser DNS over HTTPS by policy, block port 853, and lock private DNS on managed Android devices.
Browser extensions
Voice changers and text-to-speech add-ons can be installed in seconds. Allow only approved extensions on managed browsers.
Brand-new tools
A tool launched this week may not have a category yet. Blocking newly registered domains on student networks closes that gap.
Mobile data
Personal phones on mobile data are outside your network. For students, a phone policy covers this; for staff, an acceptable use policy does.
How to test the block
Use a test device set up like a real user's, signed in as a test user in the right group, with mobile data off. Repeat the key checks for each group in your policy matrix.
- Run
nslookup suno.com,nslookup udio.comandnslookup elevenlabs.iofrom a filtered client; each should return your block answer. - Open each site in two browsers and confirm the block page appears.
- Look up two or three smaller AI audio tools in the domain lookup tool, then confirm with DNS queries that the category rule blocks them.
- Open your approved speech tool and confirm it still works for the right group.
- Try to install a voice changer extension on a managed browser; it should be refused.
- Install a free VPN on a test device and confirm it cannot connect.
- On a managed phone on mobile data, confirm AI audio apps stay blocked.
- Read the logs and confirm events are tagged with the AI tools category.
Troubleshooting
Most problems after an AI category change are either over-blocking of approved tools or devices that skip your resolver.
| Symptom | Likely cause | Fix |
|---|---|---|
| Approved reading tool stopped working | It sits in the AI tools category too. | Add it to the allow list for the right groups. |
| A video app's AI voice feature still works | The feature runs on the app's own domains. | Turn it off in the app's admin console. |
| One laptop reaches blocked tools | Browser encrypted DNS, VPN or a different policy group. | Check the device's settings and group membership. |
| A new tool is reachable | Domain not yet categorized. | Block newly registered domains for student networks; check it in the lookup tool. |
| Music staff cannot reach a tool they need | Category rule applies to every staff group. | Create a group exception with a review date. |
What to tell people
A block on AI tools can feel like a block on progress. A clear explanation, and a route to approved tools, keeps people on side.
Students
AI song and voice tools are blocked on school Wi-Fi and devices. Making a fake recording of someone is never acceptable, wherever it is made.
Teachers
Explain how to request access for a lesson, and which approved speech tools are always available.
Staff in a business
Use the approved narration or music tool for company work. Never upload a colleague's or customer's voice without written consent.
Finance and help desk teams
Voices can be faked. Confirm unusual payment or access requests through a second, known channel.
Parents
The school controls its own network and devices, not phones on mobile data. Talking at home about fake audio helps.
Safeguarding lead
Filtering is one layer. Incidents involving fake audio of a student or staff member go through your safeguarding and conduct processes.
Keep up with new AI tools
No list of three domains will keep pace with the AI audio market. A dedicated AI tools category, updated every day, is what keeps the policy working month after month.
New domains are checked again after a week and a month, because many AI tools launch with a bare landing page and add features later. Filter vendors, MSPs and edtech makers license the web filtering database to run these rules in their own products.
AI music and voice blocking FAQ
Which domains do I block for AI music and voice tools?
Well-known examples include suno.com and udio.com for AI songs and elevenlabs.io for AI voice and speech. Block each with all subdomains. There are hundreds of smaller tools and new ones launch weekly, so an AI tools category rule matters more than any name list.
Why block voice cloning tools at work?
Voice cloning can be used to impersonate executives or colleagues, and uploading recordings of people's voices raises consent and data protection questions. Many businesses limit these tools to approved teams with a clear business need.
Can I block AI music tools but keep text-to-speech for accessibility?
Yes. Block the AI tools category for the relevant groups and add allow-list exceptions for the specific accessibility or text-to-speech services you have approved. Built-in screen readers and device speech features do not depend on these websites.
Does blocking these sites stop AI audio features inside other apps?
Not always. Some editing, video and office apps include AI voice or music features that run through the vendor's own domains. Manage those through the app's admin settings or your MDM, not only the network filter.
Are AI music and voice tools blocked by the CIPA categories?
Not by default. CIPA focuses on visual depictions that are obscene, child pornography or harmful to minors. AI audio tools are a policy choice for each school, usually handled through an AI tools category rule.
How does a category database keep the block current?
New AI audio tools, wrappers and mirror sites launch constantly. A dedicated AI tools category refreshed daily catches them as they are classified, without manual list updates.
AI tool categories for your filter
For filter vendors, MSPs and edtech makers: 16,000+ AI tool domains in 18 categories, inside a database of 120 million domains. Starter database tiers from $999; enterprise and OEM licences are quoted per deployment.
Suno, Udio and ElevenLabs are trademarks of their respective owners. Google, Chrome and Android are trademarks of Google LLC; Apple and iOS of Apple Inc. (iOS used under licence); Microsoft, Edge and Intune of Microsoft Corporation; other names belong to their respective owners.
These names are used here only to identify the products discussed. Web Filtering Database is not affiliated with, sponsored or endorsed by any of these companies.