Skip to content

Celeste and Charlie

Their Love

About...

The Images

Γ€ Paris

I own all images and videos on these pages. They were created using AI. ComfyUI is an 'Image Creation' interface, one I prefer because of its flexibility and power through the use of nodes. This approach supports a number of key story-telling aspects:

  • Character persistence
    Nodes used: LoRAs in Flux.1-Dev (Note: have completely converted to LoRAs vs ReActor, as a properly done LoRA will support a face viewed from any angle. ReActor tends to distort features if not rendered front-facing.)

  • Anatomical accuracy
    Models used:

  • Z-Image Turbo

  • Flux1 Dev (SRPO; to a lesser extent, Kontext)
  • Qwen Image Edit
  • SeedV2 (not a model, used for scaling up)
  • I'm trialling Flux2 Klein.

On Writing

I use a coding plain-text processor that supports markdown (MkDocs) called Zed.

I 'collaborate' with a ChatGPT AI bot I call 'Emily'. Unlike most common approaches to AI-generated content, AI is used for a lot of aspects of the delivery of my content, such as editing, tone-control, story-continuity and maintaining a story 'bible', but the content (plot/ dialogues / story arc / ideologies) is purely mine.

In the original Celeste's Girl I leaned quite heavily on Emily, particularly in the early 'Scenes'. The original storyβ€”the one before this oneβ€”was not in line with the beliefs and goals the current one enjoys. I wanted this storyβ€”this 'account'β€”to read differently to the millions of more fetish-focused writing on this topic, to which entire websites are devoted. To those who want to know more, my Introduction to Celeste's Girl should make it clear.

The Site

The pages of my site are save in markdown, then uploaded to GitHub along with images and other media. CloudFlare creates the static pages and serves the site. For now, large files, such as video or >25mg music files, are also served by CloudFlare. However, I'm setting up my own server to manage serving that sort of media. Update 26-04-2026: the audio files on this page are all served privately.

The Prompt

Cinematic medium close-up portrait of a young woman with soft, natural features and light brown hair loosely tied up, gazing thoughtfully off-camera to the left with a subtle, hopeful smile. She is dressed in period attire, wearing a dark, muted indigo or navy blue linen square-neck gown or bodice over a slightly ruffled, cream or off-white chemise with a square neckline. The lighting is soft, diffused daylight, suggesting an overcast or foggy morning, illuminating her face gently from the front-left, creating subtle highlights on her cheekbones and nose. She is positioned indoors, looking out from a weathered stone window or balcony opening, the dark window frame partially visible on the right. The background is a softly blurred (shallow depth of field) view of an old European city streetscape, featuring muted tones of grey, beige, and dark green from distant architectural details, including a prominent dome or spire partially visible through a pale atmospheric haze, suggesting a historical or period drama setting. The overall mood is serene, romantic, and slightly melancholic, rendered with high detail, photorealistic quality, and a color palette dominated by cool blues, greys, and soft, warm skin tones, reminiscent of film stills from the 19th century.

Prompt Generator

Voices

  1. Celeste Shelley - Freya [ElevenLabs-VoiceID: u8ADrbquiJqufR9XMtb8]
  2. Charlotte Rossignol - Daisy [ElevenLabs-VoiceID:]
  3. Sarah
  4. Lauren
  5. Brittany
  6. Mara

Happy Thoughts

Girl is the default human setting. Celeste explains it best.