The press lost control to AI. Today, AVALW takes it back with WratumGuard.

Today, AVALW officially launched WratumGuard. It is an advanced anti bot, anti fraud and anti abuse system, built and fully owned by AVALW. And today we already put it to work: as of today, WratumGuard blocks AI training crawlers in real time across avalw.com and all of our creators subdomains. From today, those AIs no longer see our content. Instead of articles, they receive our licensing page, live at avalw.com/ai-licensing.

WratumGuard is not just an AI blocker. It is a verdict engine that analyzes any digital request and decides, in under 10 milliseconds, whether it is legitimate. One verdict, on any request, across every surface you run. Blocking AI training crawlers is its first public mission, but its capabilities go much further.

What WratumGuard does

Today we used only one of these powers: stopping AI crawlers. And we want to be very clear, so there is no confusion: we blocked only the training crawlers. We did not touch the AIs that serve real users. This distinction is essential, and it is exactly where most of the AI versus the press debate gets confused.

Two completely different worlds of AI

Why only training? Because, of all of them, training crawlers were the worst possible scenario for us. They send us no reader, they do not cite us, they bring us no money. The only thing they do is mass download our articles and images, consuming bandwidth and server power that costs us real money, just to train the models that will appear tomorrow. In effect, we were paying for the infrastructure on which AI companies build billion dollar products.

The training AIs we blocked today

To grasp the scale of the consumption, a real example from our logs: the Meta training crawler was pulling images of hundreds of kilobytes, dozens per second, non stop. After WratumGuard went live, those exact requests now receive a 138 byte redirect instead of the 132,000 byte image.

It has nothing to do with robots.txt

The industry classic response to these crawlers has been robots.txt, a file through which a site politely asks bots not to access it. WratumGuard has absolutely nothing to do with robots.txt. It asks for nothing and relies on no one goodwill. robots.txt is a request that anyone can ignore, and some training crawlers do ignore it. WratumGuard enforces the decision at the server level, in real time, before the bot ever touches the content. It is not a no entry sign, it is a locked door.

The real problem for the whole industry

Now we move from what we did today to what this system means for the entire industry. Almost every news company in the world reports the same thing: their revenue has dropped enormously because of AI. The mechanism is simple. When a user searches or asks an AI, and the AI answers directly, the user never visits the site. No visit means no ad, no ad means no revenue. The work stays with the publisher, but the earnings stop at the AI.

WratumGuard can cover the entire spectrum, not just training crawlers. The same technology that today stops training bots can, just as easily, stop answer AIs too, the ones that steal the visit the moment the user asks. We chose to leave them free today, because they bring us visibility. But the power to stop them completely exists, and every news company could decide for itself how far to go.

This is where the economics get fascinating. If an AI can no longer see the content, it can no longer answer based on it. Without access to fresh, trustworthy information, the quality of the answers drops. An AI model is only as good as the data it sees, and the most valuable data in the world is exactly the real time news these publishers produce. Blocking AI completely is not a punishment, it is automatically a benefit for the company.

Now imagine the effect at scale. If every news company in the world used WratumGuard and cut off AI access, not only training but answer AIs too, the AI models would begin to degrade, because quality information would disappear from under their feet. At that point, the balance of power would flip: every AI company in the world would need publishers more than publishers need them.

And the only way AIs could keep their access to fresh information would be to pay a fair fee, a reasonable one, for every access. It would not be a war, simply a return to an old and healthy rule: what you take, you pay for. We believe this is the right solution. Not hatred toward AI, but a fair payment to the companies that produce the content.

And beyond AI, WratumGuard remains a complete shield. The same infrastructure that today blocks training bots can stop ad fraud, cyber attacks, fake traffic and crypto manipulation, all with a single verdict, in under 10 milliseconds, on any request. For a media company, that means not only content protected from AI, but also ad revenue protected from fake traffic and infrastructure protected from attacks.

For AVALW, WratumGuard is not just a technical feature, it is a position. We built it, we own it, and we launched it today. But its real value is for the entire press industry and beyond. It is the tool that gives companies back something they lost years ago: the right to decide who uses their work, and on what terms. In a world where AI has rewritten every rule of the press, WratumGuard is how you write your own back.

WratumGuard is powered by AVALW, at wratumguard.com. It is opening publicly soon for companies that want to try WratumGuard for themselves.

wratumguard.com
avalw.com/ai-licensing