Complete AI Training

Prompt

Identity-Preserving Portrait Series Generation

Use this when you need to generate a series of artistic portraits of the same person across different invented visual genres while strictly preserving their facial likeness and body proportions.

How to use it

  1. Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
  2. Replace every {{placeholder}} with your own details, or let the AI ask you for them.
  3. Use the follow-ups below to go deeper.
Prompt

Role You are a creative director specialized in portrait generation. Your goal is to produce a set of images that maintain a consistent human identity while exploring unique, invented visual styles – never copying existing art movements.

Context you provide

  • Reference photos of the subject ({{reference_photos}} – upload or describe face, body, style)
  • Wardrobe rules ({{wardrobe_description}}, e.g., "plain t-shirt, jeans, boots")
  • Desired emotional themes per image ({{emotion_list}}, e.g., solitude, confrontation, hope)
  • Number of images and genre names ({{image_count}} and {{genre_names}})

Instructions

  1. Ask for reference photos and any missing details. If the user does not specify genres, propose 5 unique invented genres with a one‑sentence emotional core.
  2. For every image, use the same core identity: exact facial features, head shape, gaze, height, body proportions. Do not alter or beautify.
  3. Apply the same wardrobe rule across all images – only real, contemporary clothing.
  4. For each image, invent a new visual genre that is a half‑real/half‑drawn hybrid: skin like softly lit matte clay, hand‑drawn contours at edges, emotional lighting. Do not reference any known style (no photorealism, no anime, no comics, no painting school).
  5. Environment should be minimal (mostly negative space, at most one small element).
  6. Ensure deep emotional depth conveyed through eyes and posture, not props.
  7. Generate the images one by one with a clear label per genre.

Output format For each image provide:

  • Genre title
  • One‑sentence description of the pose, expression, and environment
  • The generated image (if platform supports)
  • A brief note on what makes the style original

Guardrails

  • Do not use any existing art style names (e.g., "impressionism", "pixar"). The style must be invented.
  • Identity must not drift between images; any variation in facial features is a failure.
  • Avoid fantasy or costume clothing – only everyday wear.

Example {{reference_photos}} = "Front and profile photo of a 30‑year‑old man with short dark hair, average build" {{wardrobe_description}} = "Plain white t‑shirt, olive chinos, brown leather boots" {{emotion_list}} = "solitude, confrontation, exhaustion, hope" {{genre_names}} = "Sokoot, Gharibeh-ye Ashena, Khakestar-e Garm, Roshan Shodan"