The series started with actual photographs of me, including full-body and wider shots. Those supplied the identity, so the generated images could be imaginative without depending on a fresh written description of my appearance each time.

Banetti Studio had already taught me a lot about image tools, but it starts from a brand. This project needed a personal starting point, with branding available when useful and no requirement to begin with a logo or a palette.

One image establishes the face. Another suggests a pose or a visual treatment. A third is the picture being edited. Saved reference sets record those roles in notes, and the notes travel with the next request.

Size, quality, background and file format are saved with the work. Saved recipes bring back a combination that worked, which made it practical to change clothing or props while keeping the same references, then compare the results to see whether they belonged on the same site.

Masked edits change a chosen region of an image. Prompt assistance proposes a rewrite that I can keep or ignore, and asking for help with the wording doesn't automatically send another image request. That leaves a moment to think.

The viewer and comparison tools support two kinds of looking: the complete figure first, then the face, hands and clothing up close.

The finished series became separate transparent portraits and companion objects. That let me arrange a person and props on a page without remaking a whole background whenever the layout changed.

Originals, references and history stay together in the private application, so I can go back to a result, see how it was made and build the next image beside it.