Advanced Filtering
Block malicious domains and enforce content policy in real-time, before anything executes.
Learn more
AI Governance & Controls
Your AI policy, enforced before the connection forms
Almost every organization has written an AI policy. Far fewer have controls that make it true on devices they do not manage. This suite turns the policy into an enforcement point at the DNS lookup, where every AI service, agent, and remote MCP endpoint has to resolve a name before it can do anything. AI tools and agents operate inside the rules you set, on the devices you manage and the ones you cannot.
- 200 billion DNS queries resolved daily
- 235 million threats blocked daily
- 57,000+ organizations protected
Discovery first, then control
-
Get complete shadow AI oversight across the environment
You have an approved AI tools list, and you probably do not have full confidence it is complete. This surfaces the AI services your environment is actually reaching, per user and per device, including on hardware nobody manages and hardware nobody could install software on.
-
Control which AI tools each team can reach
Not all AI usage carries the same risk. Allow the tools you have approved, restrict the ones under review, and block the rest, with different rules for engineering, finance, and legal. The decision lands at the lookup, before a session exists and before any data moves.
-
Apply policy to AI agents as distinct entities
An autonomous agent is not a person and should not inherit a person's policy. Agents get their own profile, so you govern what a non-human identity may reach separately from the human who launched it, and you can see what it went and did once it was running.
What you get
Discover, decide, enforce, prove
-
See what your environment is actually reaching
The AI tools nobody registered, the agents nobody approved, and the remote MCP endpoints your devices resolve. Surfaced per device and per user, including on hardware you do not manage.
-
Agents as their own policy object
An agent carries its own profile rather than inheriting the policy of the person on the machine, so what a non-human identity may reach is set separately from what the human may open.
-
Enforce before the connection
Allow the AI services you have approved, restrict the ones under review, and block the rest, by user, by device, by site, and by tenant. The decision lands at the lookup, before a session exists.
-
Answer the audit question
When somebody asks which AI services your data has reached, the answer comes from a record rather than a reconstruction.
Runs on the same DNSFilter roaming client. Nothing new to deploy, and no migration attached.
Who it is for
-
For security and IT teams
One enforcement point that covers managed and unmanaged devices alike, with no new agent to deploy and no migration attached. When the board, the auditor, or your insurer asks about AI, you have a control to point at rather than a document.
-
For MSPs and partners
Your clients are starting to ask whether their people are using AI tools with company data. Most providers cannot answer, because the tools run on devices nobody manages. This surfaces them at the DNS layer, across every client, through the multi-tenant console you already use.
Make the policy real
Discover what your environment is reaching, decide what it is allowed to reach, and enforce that at the moment of the lookup.
Five suites. One engine.
Each suite is a distinct outcome of shared platform intelligence. Start with one and expand at your pace, with no replatforming between them.
-
-
Behavior Analytics
Visibility into user behavior so anomalies and indicators of compromise surface earlier and with higher fidelity.
Learn more -
AI Governance & Controls
Discover the AI tools and agents your devices reach, then apply policy to them as separate entities from the people using those machines.
Learn more -
Encryption & Privacy
Enforce privacy as policy, with encrypted transit for every user everywhere they work.
Learn more -
Threat Intelligence
Proprietary, DNS-First threat intelligence with patent-pending lookalike domain detection.
Learn more
FIND OUT MORE
FAQs
Read the answers to the most frequently asked questions. Want to know more? Visit our help desk.
AI governance is the set of controls that determine which AI tools, models, and agents may operate in your environment, what they may reach, and who may use them. In practice it spans several layers, from prompt inspection to model posture to access control. This suite handles the access control layer: Deciding at the DNS lookup whether a connection to an AI service is allowed, restricted, or blocked.
A CASB governs what a person may open, usually on a managed device and inside a session it can route. This governs what any identity, including a non-human one, may reach, before the session exists, and it works on devices no agent covers. Teams run both.
Yes. The visibility comes from DNS resolution rather than an endpoint agent, so devices that cannot run software still come into scope. That includes contractor laptops, personal phones, IoT, and guest traffic.
The device, the user, the site, and the service, plus the time. We do not read prompts, files, or responses, and we cannot tell a corporate AI account from a personal login on the same service, because both resolve the same hostname. That level of detail is still how most teams find AI usage they did not know existed.
We see remote MCP endpoints that a device resolves. We do not see MCP servers running on local or stdio transports, because those produce no external lookup.
Yes. An agent is treated as its own entity rather than inheriting the policy of the person on the same machine, so you can allow a human to use a tool while restricting what that tool may reach once it is running.
Yes. It uses the same DNS-First intelligence engine, the same console, and the same policy hierarchy as the rest of the platform. If you already point DNS resolution at DNSFilter, there is nothing new to stand up.