A Muse Made of Echoes

Introducing:
Roxanne, The Heirloom Muse

A fictional character generated using AI, blending traits inspired by vintage portraiture, classic cinema, and mythic glamour. Not based on any real individual, but a blend of multiple people whose images I’ve found to be unforgettable because of their unique features.

Roxanne by the sea.


As I begin this new chapter, I thought it was only fitting to address one of the most common questions I receive:

Who are these women I keep seeing in your AI-generated stories?

They are not real. But they are not random either.

I often begin with a question—not who is she? but what does she feel like to remember?

In my latest exploration with AI-generated portraiture, I wasn’t interested in replicating a face. I wanted to compose one. A woman who feels familiar but impossible to place. Timeless, but not ageless. As if pulled from the subconscious—half dream, half memory.

To shape her, I blended four visual threads:

  • A cinematic presence known for quiet strength and luminous symmetry.

  • A legacy figure from early screen culture—angular, poised, signature eyebrows, and unapologetically commanding.

  • A supermodel from the 90s with a strong jawline, enormous mouth and eyes.

  • And finally, a muse drawn from fin-de-siècle portraiture—delicate yet adorned, suspended in myth and ornament.

These weren’t literal references. I didn’t use names or images, but instead worked with emotional tone, aesthetic archetypes, and historic visual language.

Using Midjourney in private mode, I translated this vision into a fictional character: a woman who does not exist, yet feels as if she could have. A memory you’re sure you’ve seen before, but can’t quite place.

People have asked:
“Is that Anne Hathaway?”

Roxanne

Not entirely, Anne.

There's something in the eyes, the symmetry, the quiet magnetism that recalls her. But the truth is, she's not all Anne.

Her face and characteristics weren’t created with any single individual in mind. Rather, she’s the result of emotional and visual layering—a composite formed from qualities that felt cinematic, mythic, classic, or timeless, artistic muses, and a 90s supermodel.

There may be echoes of familiar faces, but they’ve been transformed through abstraction and recombination.

Think of her as a shadow on the wall—cast by many sources of light.


These are the seed images I’ve often used after I’ve blended four images when I created my fashion editorials.

It's important to note that even when using a chosen reference image, the AI typically replicates only about 75% of it.

Hollywood classic with a 90s supermodel look

An artist’s muse come to life.

This is where I believe AI thrives—not in imitation, but in interpretation.

The women above aren't real, but they’re real enough to feel remembered. But I know I wanted a fashion muse that was a blend of old Hollywood, Gustav Klimt’s flair, huge or odd features like noses, mouths, and eyes.

But more importantly, distinct cheekbones and signature eyebrows, as these are what would make her powerful and memorable.

I went through a dozen regenerations to finally get something that matched my original vision.


Tools Used

  • Platform: Discord (Private Server)

  • Engine: Midjourney v6

  • Images: Four images as a base but also had to describe them to prompt on Midjourney.

  • Techniques: /blend, --stylize 750, aspect ratio --ar 2:3, seed variations

  • Post-processing: None. This is a raw generation.


Disclaimer:
All visuals featured in this post were created using generative AI tools. The characters depicted are entirely fictional and are not intended to resemble or represent any real person, living or deceased.