About this blog

This is the engineering and industry blog of a team that builds a generative media product — AI image and video generation, running a catalogue of hosted models from several providers. The product is Uncutly.AI, and it generates adult content. We would rather say that plainly here than have you find it out by clicking a footer link.

That context matters for reading the posts, because it is where the numbers come from. Running an adult generative product means dealing with payment rails that most companies never touch, content rules that have to be enforced in code rather than in a policy document, and per-generation infrastructure costs that make a free tier behave nothing like a SaaS free tier. Most of what we publish is what those constraints taught us.

Where the numbers come from

Every figure in a post marked as first-party comes out of our own production systems — our model catalogue, our incident record, our search console, our cost tables. When we write that fifteen registered models were dead but still selectable, that is a specific probe we ran on a specific date. When we write that a payment channel was silently rejecting callbacks for nine days, we have the three orders that prove it.

We publish those in ratios and counts rather than in absolute revenue, spend or margin, and we do not name the vendors involved. That is a deliberate line: the shape of a failure is the useful part and can be published honestly, while our supplier relationships and unit economics are not ours alone to disclose. Where a post is reporting on the industry rather than on ourselves, it is sourced to the original and says so.

What this blog is not

• Not a funnel. There is no signup wall, no email capture, no gated download, and no post that ends by telling you to try the product.
• Not sponsored. We do not take payment for coverage, and we do not accept guest posts or link insertions. If we ever mention a vendor favourably it is because we use them and said so.
• Not a highlight reel. A large share of what is here is us being wrong in public — a global conclusion drawn from testing in one country, a search strategy that took our own traffic to zero, an anti-abuse rule that broke five hundred real accounts in a day. Those are the posts with the most reusable content in them.

Corrections

If something here is wrong, we want to know and we will fix it in place with a note saying what changed. Several posts already carry corrections to our own earlier conclusions, because we published the conclusion first and the disproof later.