AI Model Generator

Most AI model generators produce a face. Then you go back tomorrow, type something similar, and get a different person.

That is fine for a mood board and useless for a catalogue. A catalogue needs the same person in shot one and shot forty, in every garment, at every angle. The generation is the easy part; the identity is the product.

Building a model

You do not write a prompt. You fill in a builder:

• Ethnicity: 8 options
• Gender: 3
• Size and body: petite, slim, athletic, average, curvy, muscular, plus-size, XXL
• Age range: teen (16–19) through senior (60+)
• Skin tone: 6
• Hair: length, style, colour
• Expression: 5
• Height: 3

Plus up to 600 characters of free direction for anything the fields do not cover, and an optional starting photograph if you have a reference you want the model to resemble.

Why the view set matters

The builder does not stop at a portrait. It runs through a short pipeline:

1. Candidates: several interpretations of your attributes, so you can choose rather than accept.
2. View set: front, side and half-body renders of the chosen candidate.
3. Ready: the model is saved to your library.

That view set is the thing that makes identity hold. Every later generation references it, which is why the model does not drift between a hero shot and a detail shot, or between a studio backdrop and a street scene.

Reuse is the whole point

Once a model is ready it behaves like any other asset in your team library. Any teammate can pull it into a shoot. It costs nothing to keep, and building a second one, at a different age or size or for a different market, costs the same as the first.

That changes a decision most brands make by budget rather than by intent. Showing a garment on more than one body type stops being a line item and becomes a checkbox.

What it does not do

It does not put a real person's likeness into your catalogue. Models are generated from attributes, not from scraped photographs of identifiable people. If you start from an uploaded reference, that is your own image and your own responsibility to have rights to.

Every generated image also exports with a signed C2PA manifest and an IPTC DigitalSourceType label marking it as AI-generated, so the disclosure travels with the file wherever it is published.