AI can now generate what filters used to block. Here is exactly how we handle that.
For years, blocking pornography meant blocking a list of known adult sites. That era is over. Thousands of AI image and video generators have launched in the last two years, and a large share of them will produce explicit material on demand. This page explains what we block, how, and where the honest limits are.
What changed.
The problem is worse than "more sites," in three specific ways.
It is generated, not searched. There is no fixed page to block. A prompt makes something new every time, so a filter that only knows yesterday's URLs is always behind.
It hides in plain sight. Many of these tools are marketed as general "AI art" or "AI companion" apps, and content filters categorize them as technology, not adult. A normal porn filter waves them straight through. That is exactly how a well-known generator slipped past our own filter, until we caught it and rebuilt the approach this page describes.
"Nudify" and deepfake tools cause real harm. A category of apps exists only to fabricate nude images of real people from ordinary photos. They have been used to target and harass women and minors, and they are the subject of active lawsuits by public officials. Blocking them is not about temptation. It is about refusing to let that harm run on your network.
1. We block by capability, not by category.
Our default rule is simple: if a site can be used to generate not-safe-for-work material, we block it by default. We do not wait for a site to label itself adult, and we include the dual-use tools that merely permit it. A tool that can be pointed at darkness does not get the benefit of the doubt on this network.
2. You control how much we block.
Blocking by capability is deliberately broad, so the strength is yours to choose. A standard tier covers every NSFW-capable generator. A stricter tier adds the social platforms where explicit material circulates openly. A maximum tier goes further still. And you can always make an exception for a specific site you legitimately need, or add your own sites to the wall. This is your covenant, not our cage.
3. The list updates itself, every week.
New generators launch constantly. An automated job refreshes our blocklist every week from maintained community sources, and promotes newly discovered nudify and deepfake domains the moment they surface. On top of that we run periodic deep research sweeps that hunt specifically for new AI image and video tools. This list is never "finished," and we will not pretend it is.
4. DNS is one layer. We will not pretend it is the whole answer.
A DNS filter blocks websites. It cannot see inside a phone app that is already installed, and it cannot stop a new app being downloaded from an app store. That is a real limitation, and hiding it would make us like every tool we replaced. So we close the gap at the operating system instead: our device lockdown guide shows you how to restrict apps by maturity rating on iPhone and Android, without the miserable "block the entire App Store" approach.
5. Accountability, not just a wall.
The point is not a perfect cage. A lock you can open yourself is only a reminder. The safeguard is a person: a partner who holds the passcode, who gets notified, and who unblocks nothing silently. Every layer on this page assumes that person exists. That is the difference between a filter and a covenant.
What we cannot promise.
We cannot catch every site the moment it launches. We catch most, fast, and we keep adding.
We cannot stop a determined person who has full admin control of his own device. What we can do is remove the easy paths, make the hard paths visible to an accountability partner, and hand you the honest tools to lock a device down properly.