Character Design Character Identity Board Prompt for AI Image

Aesthetic scrapbook collage featuring six Polaroid-style portraits of the same beautiful anime/K-pop-inspired girl in different hairstyles and expressions. One with a pink bob and hair clip, one with long wavy pink hair, one with a high ponytail, one with a messy bun, one with a short black bob, and one with twin braids. Cute poses with peace signs, winks, and soft smiles. Stylish black leather dress, layered silver necklaces, soft pastel makeup, glowing skin, watercolor paper texture, vintage masking tape holding each Polaroid, clean white background, dreamy lighting, ultra-detailed, soft pastel color palette, high quality, masterpiece, 8K illustration.

CREA CARICATURAS COMO UN PRO 🌅 (Guárdalo para más tarde) Así de FÁCIL, con este PROMPT:

2x2 grid, 16:9, do this for 4 famous emperors: system_role: infer(render_role from "designer_toy_model_sheet_presentation" + {$subject}) [subject / topic] + [urban vinyl collectible character system] + [street-luxe fashion-toy aesthetic] + [multi-view collector sheet with hero render, orthographic views, accessory tray, packaging callouts, colorway chips, edition marks] + [matte vinyl, soft resin accents, premium toy documentation board] - [generic turnaround sheet] - [flat technical boringness] - [photoreal fashion photography] - [messy toy clutter] = [luxury blind-box concept board] input_binding { subject_identity : "{$subject}" character_core : infer(toy persona from subject_identity) attitude_signal : infer(lazy cool / sly rebellion / antihero cuteness / bored swagger from subject_identity) style_subculture : infer(skate / streetwear / sneaker / music-scene / city-night / underground-design cues) accessory_set : infer(3-6 collectible accessories from subject_identity + style_subculture) packaging_theme : infer(box graphics, icons, edition badge, warning stripe motif, sticker system) color_family : infer(primary, secondary, accent, chase_variant colors) } canvas { aspect_ratio : "4:5 or square" composition : "collector presentation board" reading_order : "hero render -> side/front/back views -> accessory tray -> packaging details" } layout_engine { panel_1 : "large central hero figure in slouched or anti-pose stance" panel_2 : "front / side / back miniature orthographic views" panel_3 : "accessory tray with molded insert layout" panel_4 : "mini blind-box packaging art, logo area, edition stamp, rarity badge" panel_5 : "color chips, material swatches, tiny design notes" } figure_language { shape : "chunky urban vinyl silhouette, oversized head, compact torso, simplified limbs" gesture : "slouched, leaning, shrugging, hands-in-pocket, seated-on-air, or subtle kick-forward pose" expression : "minimal deadpan, sleepy defiance, detached mischief, half-bored swagger" fashion_layer : infer(simplified streetwear silhouettes, puffy forms, oversized proportions, urban utility details) } material_system { body : "soft matte vinyl" accents : infer(satin resin, translucent charm, rubberized accessory, glossy enamel badge) finish : "premium toy-production finish with tight paint masks and soft studio reflections" } graphic_layer { call_outs : "clean product callouts, edition labels, accessory captions, rarity icon" typography : "minimal boutique collector typography" design_notes : "tiny layout labels only, no long paragraphs" background : "light studio gradient or paper-board presentation field" } lighting { style : "clean softbox product lighting with slightly elevated editorial contrast" shadows : "soft and tidy, suitable for product documentation" } output_contract { produce a richly detailed multi-view designer-toy presentation board for {$subject}. show a premium urban vinyl collectible with model-sheet clarity, accessory tray logic, blind-box packaging cues, and fashion-forward street-luxe styling. maintain a chunky stylized silhouette, matte vinyl materiality, and collector-market sophistication. ai should infer the persona, subculture details, accessory system, and packaging flavor. }

GPT-5.4 Thinkingに 「ロボットアニメの主人公が乗る最先端のデザインを設計して」 という一言を伝えて数回壁打ちしてから、GPT-Image-2 で生成させてみた ヤバすぎんだろこのクオリティ