Lock a Character From One Reference With This ChatGPT Model Sheet Prompt
From M. Asif (@meAsifAi): upload one reference to ChatGPT image 2.5, lock identity, and generate a production character model sheet.
This workflow comes from original author M. Asif (@meAsifAi), published on X on September 10, 2026. It was created with ChatGPT image 2.5: upload a single reference image, lock the character's identity with an Identity Lock prompt, and generate a production character model sheet.
Image source: M. Asif (@meAsifAi) on X
The point is not a pretty reference picture but a visual source of truth for everything that follows. According to the author, a structured reference sheet keeps the character stable when poses, angles, and expressions change, and works as a foundation for video pipelines where identity and continuity must stay consistent.
The original prompt and the Identity Lock block
The verified opening of the author's prompt is as follows:
Create a premium professional character design reference sheet / production model sheet based strictly on the provided reference image.
The REFERENCE & IDENTITY LOCK block then pins the uploaded reference as the single source of truth for the character's identity. It asks the model to preserve the exact facial identity, facial structure, hairstyle, hairline, eye shape and color, eyebrows, nose, lips, skin tone, body proportions, physique, age appearance, distinctive features, costume design, accessories, footwear, colors, patterns, and all recognizable visual details. It also instructs the model not to redesign, beautify, simplify, age, de-age, or reinterpret the character.
Use the uploaded reference as the single source of truth for the character's identity. Preserve the exact facial identity, facial structure, hairstyle, hairline, eye shape and color, eyebrows, nose, lips, skin tone, body proportions, physique, age appearance, distinctive features, costume design, accessories, footwear, colors, patterns and all recognizable visual details. Do not redesign, beautify, simplify, age, de-age, or reinterpret the character.
The prompt also includes a CHARACTER DESIGN ANALYSIS step that analyzes the character design internally before constructing the sheet. Only the beginning of that step was verified in the desk memo, so this article does not reconstruct its details.
How to use it and what keeps continuity
Per the author's follow-up notes, the flow is simple. Upload one reference image to the ChatGPT image model and generate a model sheet covering front, side, and back views plus expressions and poses with the Identity Lock prompt above. It is described as a production workflow, not just a pretty reference: the structured sheet serves as a consistent visual foundation while individual cuts get piece-by-piece refinement afterward.
For continuity, the author points to recurring visual anchors. When checkable elements such as the costume, distinctive details like freckles, and footwear stay identical across views, the character remains recognizable. With the identity locked, different poses, angles, and expressions are possible while the core of the character stays anchored.
Verified reception and one caveat
Replies confirm the identity-lock trick works and produces production-ready, consistent output across angles. One dissenting craft note was also verified: an extremely detailed sheet can push ChatGPT toward defaults, so refining piece by piece may work better. The safe combination is to lock the foundation with the reference sheet first, then refine only the cuts that need it.
Original source
- Original post by M. Asif (@meAsifAi): https://x.com/meAsifAi/status/2097949170594193427