Our Services

There is no easy answer for that question. I’ve seen people train amazing loras with 30 pictures and I’ve seen some AI model makers use 600+ images. I think the answer really depends on what your end goal is. I’ll just say this … it’s easier to train a simple face lora than it is a full blown fine-tuned model.

My. Our license only covers the AI portion of the training. That means you can use the original images to train an AI model but you can’t publish the original images. That is not covered in your license.

Yes. Your license does not limit how many times you use the dataset (the photos) to train AI models. It only states that you are the only person allowed to do it and the license isn’t transferrable. This means that we only give YOU (or your company) permission to use these images. You can’t share them with others nor can you sell your license to someone else. If someone you know wants to train their own AI model with our images then they’ll need to come to us to purchase a AI training dataset license of their own.

No we only license the datasets. We don’t do the actual training. 

No I’m sorry. All datasets are pre-defined and come how they are. Some momdel sinclude 100 photos, some include 500 (or more). We can’t break up sets. They come how they come. It’s up to you to decide which images you want to use from the datasets (if not all) when you are training your own AI models.

Of course. What you create with your AI model is your own work and you can name her whatever you so desire. Your AI model is no longer the person from the original photographs, she’s a derivative work. The derivative work becomes a second, separate work independent from the first.

Absolutely. If you violate the terms of your license, then your license becomes invalid, meaning that you are open and subject to a variety of law violations both in the realm of copyright, revenge porn as well as the new AI regulations.

These terms are spelled out in your license, but basically …

• You are not allowed to use the data to generate images that infringe, misappropriate, or violate any third party’s intellectual property or other rights or generate spam or false, misleading, deceptive, harmful, or violent imagery.
• You cannot use the dataset to generate images depicting an underage person. 
• You are not allowed to use the dataset to generate images that appear to be real people from the original images.
• You are not allowed to use the data to generate images that would break the law. This includes but is not limited to fraud, deception, or impersonation of an individual with the intent to deceive.
• You are not allowed to use the data to generate deepfakes, which are defined as taking the body of an AI image generated from our dataset and placing the face of a celebrity or other person on it who does legally consent to perform in that type of content in the first place. CONSENT MATTERS.

Our model releases are signed based on certain agreements, and as someone licensing the datasets based on these releases, you are subject to these agreements and terms as well.

All models were 18 years of age or older when the images were taken. We have all associated 2257 documents as required by law. In addition, all AI datasets (photosets) come with signed model releases where the model consented to take the images and has signed the appropriate rights that allow us to use the images in this way. Not all models sign away this right but in our case, all of our photosets allow for deriviate works.

Typically speaking, fine-tuned AI models based on the likeness of a person is considered a derivative work, which is a work based on or derived from one or more already existing works. In other words, our AI models are trained on “work” previously done by models, content creators, and/or adult performers, who consented to perform in and sign the required model releases. All content that our company has legally licensed and rights to use.

A “derivative work” is a work based upon one or more preexisting works. Only the copyright owner has the right to make or authorize an adaptation of the original work.

You can’t just take random photos you find off of the Internet and use them without the permission of the owner of that content. That’s like me going to your house and taking your car without permission. I mean, I only “borrowed it,” so what’s the big deal, right? Wrong! Taking something you down own without permission is theft. Every parent teaches their child that at an early age, and that includes photos and videos.

Even when it comes to making AI based on those “borrowed” images. Copyright law is brutal and can cost you as much as $250,000 per violation. That means if you borrowed 15 photos to make a Lora, that’s $250,000 * 15. That’s almost $4 million … for only 15 images. It takes more than that to train a decent lora. And that’s only if the judge decides to go easy and not count every generated image from that AI model as a violation. Assuming the judge goes easy on you and only punishes you for the original 15 images, you are still out millions, and that doesn’t even include the cost of hiring a lawyer to defend yourself in the first place. Don’t think you won’t get caught either because lawyers are gearing up to take on these cases as we speak. And all 5 of the top 5 biggest adult companies now have legal teams dedicated to this sort of litigation.