Anissa shares a GPT Image 2 prompt for editorial portraits with fabric-cut typography

On June 24, Anissa (@SimplyAnnisa) posted on X a prompt for GPT Image 2 on ChatGPT aimed at editorial beauty posters. The idea: a vertical portrait with the face framed behind crumpled fabric, while a giant condensed typeface runs through the fabric spelling out a word in stacked syllables—with eyes, nose, and lips visible between the gaps in the letters—plus a vertical side word, a barcode block, and a muted fashion magazine palette.

AI Prompt · How to do it

Second result of the same prompt with different variables. @SimplyAnnisa

Steps

1. Open chatgpt.com with GPT Image 2.
2. Replaces the six prompt variables (character, hair, cloth, split word, side word and palette).
3. Paste the entire prompt; ask for 4:5 format if it does not respect it.
4. The thread includes four examples with different combinations.

Prompt to copy

Vertical editorial poster portrait of [CHARACTER] with [HAIR], face tightly framed behind crumpled [FABRIC], oversized bold condensed sans-serif typography cut through the fabric spelling "[SPLIT]" in stacked syllables across the whole image, eyes, nose and lips visible through the letter openings, small vertical side word "[SIDE]", tiny barcode block near the lower right, high-fashion beauty lighting, realistic skin texture, soft shadows in the folds, muted [PALETTE], clean magazine-cover composition, sharp typography hierarchy, premium print finish, no extra words, no messy layout, no cartoon style, no 3D render, ar 4:5

Text in English according to the original publication. Anissa credits the creation to GPT Image 2 via @TopviewAIhq.

What makes this prompt special

It combines editorial beauty photography with typography as a sculptural element: the word does not superimpose the portrait, but rather runs through it as if the fabric were cut following the letters. The visual trick—eyes, nose, and lips peeking out of the typographic openings—requires precision in the prompt; Without that phrase, models often place the text on top of the face or distort the features.

The final restrictions (no extra words, no messy layout, no cartoon style, no 3D render) point to a printed magazine cover finish, not a digital collage or 3D render. The barcode block and the side word [SIDE] reinforce the aesthetics of packaging or limited edition.

How to fill variables

Try [CHARACTER] with specific descriptions (young woman with freckles, androgynous model with sharp cheekbones). In [HAIR], specify texture and color (slicked-back dark wet hair, loose copper curls). [FABRIC] can be ivory silk, crinkled organza or translucent mesh. Split [SPLIT] into syllables that fit the layout (BEAU-TY, VOG-UE, ELLE). [SIDE] is usually a short vertical word (ISSUE, NO.07, PARIS). In [PALETTE], name muted tones (dusty rose and charcoal, sage green and cream).

Anissa had already made other GPT Image 2 prompts viral in June — such as transport pass and portrait with gradient in Nano Banana 2—; This piece aims at the beauty editorial poster niche. The thread exceeded 2,400 views in a few hours.

Before publishing

AI-generated barcodes are decorative and non-functional. Respect image rights if you use references from real people. Review OpenAI's policies for generated content.