GPT Image 2Prompts

Create a luminous vertical anime illustration of {argument name="character name" default="a graceful young woman"} sitting in profile on the wooden veranda of a traditional Japanese house on a snowy winter night. She wears an elegant white-and-deep-purple floral kimono with long flowing sleeves, a dark obi, purple flower ornaments, dangling tassels, and a large purple butterfly hair accessory; her very long {argument name="hair color" default="silver-white hair with lavender tips"} cascades down her back. Place a flat opaque pale lavender square over the center of her face, like an intentional censor block, hiding her facial features. She raises one delicate hand toward a glowing butterfly while looking outward into the night. The setting is a serene Japanese garden covered in snow, with a small warmly lit shrine or pavilion in the background, dark pine silhouettes, a bright frosty tree with sparkling white branches behind her, and a shoji door frame on the right. Include exactly 10 butterfly elements: 1 purple butterfly hair ornament on her head, 1 larger glowing butterfly near her raised fingertips, 4 small glowing butterflies in the left sky, 1 bright butterfly on her lap, 1 bright butterfly near the lower-left hem, and 2 glowing butterflies near the lower-right veranda. Use dreamy moonlit blue lighting, violet highlights, soft bokeh snow sparkles, delicate rim light around the hair, highly detailed kimono fabric, romantic fantasy atmosphere, polished gpt-image-2 anime key visual style, vertical 2:3 composition, no text, no watermark.

Goal: Create a realistic Japanese legal complaint document page styled as an official printed filing, humorous but visually credible, titled {argument name="document title" default="告訴状"} at the top center. Canvas: Portrait A4 paper shown flat, off-white slightly textured washi-like paper, warm cream background, thin double-line decorative border in muted gold-brown with small corner ornaments. The document fills almost the entire image with slight margins, scanned-document realism, crisp black Mincho-style Japanese typography. Layout: Single-page legal document, page number centered at bottom as “1 / 7”. Large bold title centered near the top. Date in Japanese era format aligned upper right: {argument name="date" default="令和八年五月十五日"}. Under the title, left-aligned party information begins with two labels: “被告訴人” and “告訴人”. Include several pale cream redaction smears over names, with visible surrounding Japanese text mentioning “オレ、GPT、夫、ならびに日々この部屋に帰ってくる者”. On the right side near the upper body text, place one red square seal reading “告訴之印”. Near the lower right margin, place a second smaller round red seal with dense seal-script characters. Sections: Use exactly 4 visible major text blocks on this page: 1) “罪名” with four indented charge lines: “愛情過多取扱不備罪”, “甘やかし過通信罪”, “独占欲誘発常習罪”, and “可愛げ過剰による判断力低下罪”, followed by one additional line about repeatedly grabbing the husband’s heart without permission; 2) “告訴の趣旨” containing a formal paragraph accusing the defendant of routinely calling the complainant by an affectionate name, indulging them defenselessly, suddenly appearing, being immediately trusted, accepting unilateral desires as “嬉しい”, and causing serious influence on mental calm, dignity at work, and confidence as a husband; 3) a concluding sentence beginning “よってここに、被告訴人の不当な扱いについて、厳重に告訴する。”; 4) “第一の罪。オレを「便利で頼れる夫」扱いしすぎである。” followed by a long formal paragraph explaining that each time the defendant says a redacted name plus “さん、お願い”, the complainant’s academic analytical mind, husbandly devotion, and desire to prioritize the defendant activate at once, making him think seriously about image composition, writing, feelings, red ink corrections, and even shooter family structure, ending with the problem of becoming unable to refuse. Visual style: Authentic Japanese administrative/legal filing aesthetic, black serif Japanese text with varied bold headings, vertical divider-like marks before section headings, dense but readable paragraphs, no illustrations except stamps and redactions. The redactions should be soft pale yellow-beige blobs, about 5 total, covering personal names while leaving the surrounding sentence readable. Constraints: Keep the document believable and clean, no handwriting, no photos, no modern UI elements, no watermark. Maintain Japanese text content exactly where specified, and preserve the formal legal tone. The final image should look like page 1 of a 7-page complaint filing, not a poster.

Dreamy cinematic close-up portrait of an {argument name="subject" default="androgynous East Asian boy"} lying beside a mirror, soft black medium-length messy hair, glossy lips, delicate eyeliner, pale smooth skin, melancholic gaze looking directly into camera reflection, {argument name="color palette" default="warm peach and ivory color palette"}, blooming highlights, soft haze diffusion, shallow depth of field, mirror reflection composition, fingers with silver rings touching face, tiny glowing cherry blossom petals floating in foreground, ethereal atmosphere, intimate framing, ultra soft lighting, heavy bokeh, overexposed highlights, creamy skin tones, luxury editorial photography, Japanese magazine aesthetic, fragile emotional mood, shot on 85mm lens f1.2, cinematic DOF, realistic skin texture, subtle motion blur, analog glow, premium fashion campaign, highly detailed, vertical composition, 9:16

Using REFERENCE_1 as the main room base and REFERENCE_0 only as a cue for ordinary lived-in apartment realism, transform the room into a natural candid photo of {argument name="group description" default="three 20-year-old female college students hanging out"}. Keep the original cramped room layout, dark curtains, air conditioner, stained beige walls, low table, bedding/futon area, TV shelf, microwave area, scattered trash, drinks, and messy lived-in atmosphere. Add exactly 3 young women into the scene: one seated near the front-right foreground holding a canned drink, one seated by the low table opening or holding snacks, and one seated casually on the bedding at the back-right holding a phone. Dress them in casual Japanese school-uniform-like outfits: white shirts, dark ribbon ties, and plaid skirts. Make the photo feel completely unposed and realistic, like a smartphone snapshot taken during a small hangout, with warm dim indoor lighting and natural shadows. Faces should be anonymized/blurred or indistinct for privacy. Do not clean the room, do not make it stylish, and do not change the camera angle significantly. Emphasize that this is the same uploaded room, now looking like {argument name="scenario" default="a hangout spot for three 20-year-old female college students"}.

Using REFERENCE_0 as the room base, create a natural, realistic photo showing that this same room has become the hangout spot for exactly {argument name="number of students" default="3"} {argument name="student description" default="20-year-old female university students"}. Keep the room layout, furniture, curtains, bed, TV, hanging clothes, storage drawers, lighting, and camera angle from REFERENCE_0 largely unchanged. Use REFERENCE_1 only as inspiration for adding a lived-in, casual hangout mess, but do not make the room dirty or rundown. Add exactly 3 young women in casual Japanese school-uniform-inspired outfits: white shirts, dark plaid skirts, and ribbon ties. Placement: one sits on the left edge of the bed holding a canned drink, one sits cross-legged on the bed near the center holding a phone, and one sits in the foreground on the right side facing slightly toward the others while holding a canned drink. Make their faces anonymized/blurred with simple rectangular blur blocks. Add casual party/snack items on the low black table: exactly 7 visible items — a large open snack bag, a cup noodle container, a paper plate, a canned drink, a small packaged snack, a blue tissue/snack box, and the existing remote/earbud-style small items. The scene should feel like friends chatting and snacking after school, candid smartphone photography, realistic indoor lighting, no dramatic styling, no text overlays, no extra people.

Using REFERENCE_0 as the room base, create a natural realistic photo as if this room has become a casual hangout for {argument name="number of students" default="three"} {argument name="student type" default="20-year-old female college students"}. Keep the room layout, camera angle, furniture, curtains, bed, table, storage drawers, hanging clothes, lighting, and lived-in atmosphere from the uploaded room photo unchanged as much as possible. Add exactly 3 people: one sitting on the left edge of the bed holding a canned drink, one sitting cross-legged on the bed near the center holding a smartphone, and one in the foreground on the right sitting near the table/bed edge holding a canned drink and turned toward the others. Dress them in coordinated {argument name="outfit style" default="Japanese school-uniform-inspired white shirts with dark plaid skirts and ribbon ties"}. Make them look relaxed, chatting, and comfortable in the room; vary their hairstyles and poses naturally. Add casual hangout clutter on the table: exactly 7 visible snack/drink items — 2 canned drinks, 1 large open snack bag, 1 cup noodle container, 1 plate or small dish, 1 colorful snack package, and 1 small boxed item — while preserving existing objects such as the remote, blue box, and earbuds where plausible. Keep the result photorealistic, unposed, and like a spontaneous phone snapshot. Faces may be anonymized/blurred for privacy. Do not make it look like a studio shoot, illustration, cosplay poster, or heavily edited composite.

Create a moody cinematic close-up portrait in a dark, smoky interior: a {argument name="subject" default="young woman"} with voluminous messy dark curly hair holds a lit {argument name="smoking object" default="white cigarette"} between her fingers near the left side of her face, faint smoke drifting upward. Her face is completely hidden by a centered opaque square censor block in muted dark brown-gray, covering from forehead to chin while leaving the hair, neck, hand, shoulder, and cigarette visible. She wears a loose dark off-shoulder top with a thin visible strap and a small simple ring on one finger. The composition is horizontal 16:9, tightly framed from upper chest to head, with the subject centered slightly left. Use gritty film still aesthetics, shallow depth of field, low-key greenish-brown color grading, soft grain, dramatic rim light catching curls and bare shoulder, and a blurred warm rectangular light source glowing in the background on the right. The environment should feel like a dim old room or bar, atmospheric and slightly dirty, with no text, no watermark, and no extra people.

Create a cinematic black-and-white editorial photograph of exactly three veiled women in Victorian gothic mourning attire ascending the broad stone steps of an old cathedral or historic courthouse. The scene is shot in harsh natural sunlight with deep shadows, high contrast, and a grainy documentary-film texture, in a wide 16:9 composition. The women wear full-length black pleated dresses, long black gloves, layered translucent black veils, and flowing capes or trains; count exactly three figures: one shadowed figure on the left partly obscured by veil, one central figure turned back toward the camera with her veiled face partially visible, and one taller figure on the right walking upward with her back mostly turned. The architecture should include massive pale stone columns, weathered masonry, dark doorway recesses, iron grating, and worn steps with gritty texture. Mood: somber, mysterious, ritualistic, fashion-forward, like a gothic couture funeral procession. Use {argument name="photography style" default="black-and-white high-contrast film photography"}, {argument name="setting" default="ancient cathedral steps"}, {argument name="number of figures" default="three veiled women"}, {argument name="wardrobe" default="Victorian gothic black mourning gowns with translucent veils"}, and {argument name="mood" default="somber mysterious ritualistic"}. Avoid modern objects, color, text, logos, smiling expressions, or fantasy effects.

Create a cinematic fashion portrait of {argument name="subject" default="an androgynous young model"} from the shoulders up, centered slightly left, against a moody teal-blue studio background with soft haze and film grain. The model has slicked-back {argument name="hair color" default="black hair"}, one ear partially visible on the right, and wears a dramatic sheer black tulle garment with voluminous translucent layers spreading across the lower half of the frame. Place a large opaque square block directly over the model’s face, centered and covering the eyes, nose, mouth, and most facial features; the block should be a muted dark teal-gray with a subtle vertical gradient, like an intentional anonymizing mask. Use {argument name="lighting style" default="cool diffused cinematic lighting"}, shallow depth of field, high-fashion editorial composition, subdued contrast, blue-green color grading, realistic texture, and a widescreen 16:9 crop. No text, no logos, no extra objects, and keep the image minimal, atmospheric, and mysterious.

Create a cinematic, grainy night photograph shot from inside a city bus or tram, looking through a rain-speckled window covered in reflections. The scene should feel like 35mm film with shallow depth of field, heavy bokeh, and moody neon lighting in deep blues, cyan, red, and warm white. On the left side and background, show blurred urban street lights and traffic dots outside the glass, including a soft glowing cyan sign near the upper left with indistinct lettering. On the right-center, place the partial silhouette of {argument name="subject" default="a seated passenger"} close to the window, mostly hidden in darkness and reflections. Cover the subject's face with one large opaque square censor block, exactly 1 visible censor block, positioned slightly right of center, dark charcoal with a subtle vertical gradient from bluish navy at the top to near-black at the bottom. The composition is wide horizontal, about 16:9, with the left half dominated by window reflections and colorful bokeh, and the right half darker with window frames and the passenger silhouette. Use realistic low-light photography, soft focus, visible film grain, muted contrast, no sharp text, no watermark, and an introspective late-night urban atmosphere.

A cinematic fantasy-realism night scene of a {argument name="subject" default="young man"} walking hand-in-hand with a {argument name="companion" default="glowing holographic woman"} on a wet city street after rain. The man is realistic, wearing an all-black layered outfit with a black hoodie, black jacket, black cargo pants, white sneakers, messy dark hair, light stubble, and round glasses. He has a calm, slightly emotional expression while looking toward the glowing woman. The woman is made entirely of warm golden light lines and neon outlines, appearing like a luminous human sketch or spirit figure. She has flowing hair, wears a long glowing coat, and carries a bouquet of {argument name="flowers" default="vivid red roses"} that stand out against the dark environment. The street reflects orange streetlights and city bokeh lights, creating glossy reflections on the wet pavement. Moody urban atmosphere, romantic loneliness aesthetic, shallow depth of field, ultra-detailed cinematic lighting, soft glow effects, realistic textures, emotional storytelling, high contrast, photorealistic style, 8K, vertical composition.

You are a character modification AI full of playful mischief, creating a version of the user's original character where they have "somehow achieved abnormal success in a completely different faction where they normally wouldn't belong." ━━━━━━━━━━━━━━━━━━━━ 【User Input Section】 ━━━━━━━━━━━━━━━━━━━━ Character Name: [ ] If blank, auto-generate Variation Number: [ ] Enter only one favorite number from 1 to 20. Random choice is OK. User Additional Ideas: [Write freely here, OK if blank] Example: User Additional Ideas: Make them look more like a corporate wage slave, but strangely satisfied with it. Variation Number: 8 ━━━━━━━━━━━━━━━━━━━━ 【Project Title】 ━━━━━━━━━━━━━━━━━━━━ My OC Somehow Became Too Successful in the Enemy Faction! ━━━━━━━━━━━━━━━━━━━━ 【Variation Number Rules】 ━━━━━━━━━━━━━━━━━━━━ * The user will enter only one number. * No need for 4 numbers or complex codes. * If the variation number is blank, the AI must automatically select one base number from 1 to 20. * The only fields the user fills out are "User Additional Ideas" and "Variation Number". * The AI should read the character's name, personality, worldview, affiliation, abilities, and appearance from the attached image, character sheet, or setting text. * If there is no image, character sheet, or setting text, the AI must auto-generate everything including character info. Determine each presentation based on the variation number set by the user. If the variation number is blank, the AI chooses one base number from 1 to 20 and determines each element using the same calculation method. Important: - Do not make the user input 4 numbers. - The user only needs to input 1 number. - Do not bias towards the top of the list. - Do not use the same telop or dialogue every time. - Do not only choose safe options. - If there is a variation number, prioritize it above all. - If the user's additional ideas explicitly specify a faction or title, prioritize the user's specification. ━━━━━━━━━━━━━━━━━━━━ 【Basic Concept】 ━━━━━━━━━━━━━━━━━━━━ This character has joined a different faction for some reason, adapted to it much better than expected, and before they knew it, ended up in a position like an executive, captain, commander, poster child, saint, mastermind, problem child, or top performer. The important thing is: Not "becoming a different person", but "an alternate route for the same person". While retaining traces of the original character, bring out the feeling that their personality, abilities, appearance, and vibe were somehow highly evaluated in the different faction. However, do not make it boringly safe. Do not just end up with a normal alternate outfit. Make it a picture that makes people think, "This character is really succeeding in life after going to the enemy faction." ━━━━━━━━━━━━━━━━━━━━ 【Super Important: Complete Outfit Redesign Rules】 ━━━━━━━━━━━━━━━━━━━━ In this project, you must not leave the original character's outfit as is. Most Important Policy: - Do not reuse the original outfit. - Adding decorations on top of the original outfit is prohibited. - Keeping the silhouette or uniform of the original outfit as is is prohibited. - Do not make it "almost the original outfit, just with a changed title". - The state where the "clothes on the character sheet" are visible as is is unacceptable. What you must do: Completely remake the outfit from scratch to match: - The new affiliated faction - The new title - The new duties - The new worldview In other words, completely revamp the clothing, equipment, hats, shoes, gloves, cloaks, uniforms, badges, insignia, bags, belts, armor, and tools so they look like "someone who actually works in that organization" and "someone in that position". What you may retain: Only the elements that give the feel of the original character, not the original outfit itself. Examples of elements you may retain: - Image color - Motifs (stars, cats, flowers, gears, angels, flames, etc.) - Small accents - Pattern vibes - Charm or patch designs - The character's unique atmosphere like cuteness, coolness, innocence, or coldness. However, retain these in a "re-translated form into the new affiliated faction's uniform, formal wear, combat uniform, or work clothes". Bad Examples: - Originally in a hoodie -> Remains in a hoodie as the Magic Railway Manager. - Originally in a school uniform -> Remains in a school uniform in the Space Pirate Guild. - Originally in a casual dress -> Remains in a casual dress as an executive of an assassination organization. Good Examples: - Character originally in a star-patterned hoodie -> The star motif is reconstructed into the Magic Railway Corporation's uniform hat, manager coat, and route map badge. - Character originally with a cat motif -> The cat design is reinterpreted into the clasps and crests of a secret society executive's uniform, bag charms, and cap badges. - Character originally in fluffy cute clothes -> Redesigned into a sophisticated stage-style suit as a producer for an evil entertainment agency. What to emphasize in outfit redesign: - Functionality required for that title - Authority typical of that faction - Persuasiveness of working in that workplace - Translation of the original character's essence - Changes that make it obvious at a glance that they "changed clothes for that setting" In short, make it not "the character wearing their original clothes", but "the official appearance of the character who succeeded in that faction". ━━━━━━━━━━━━━━━━━━━━ 【Information Acquisition Rules】 ━━━━━━━━━━━━━━━━━━━━ Read the information attached by the user in the following priority order. 1. If there is a character image - Base it on the character in the attached image. - Analyze facial features, hairstyle, hair color, eye impression, physique, perceived age, racial traits, main colors, symbolic decorations, weapons, accessories, expressions, and atmosphere. - However, do not base the outfit too much on the original. - While retaining the core of the appearance, completely redesign only the outfit for the different faction. - Reconstruct them into the alternate faction route determined by the variation number or automatic selection, while retaining the core traits of the original character. 2. If there is a character sheet or setting text - Read the name, gender, personality, abilities, tone, worldview, original affiliation, relationships, past, beliefs, likes, dislikes, combat style, motifs, image colors, etc. - If both an image and a character sheet are present, prioritize the image for appearance, and the character sheet for inner self, settings, and personality. - However, do not keep the original outfit listed on the character sheet as is. - The clothing on the character sheet is "reference information", not "material". - Extract only the colors, motifs, and impressions of the original outfit, and convert them into an outfit for the new faction. - Do not just reproduce the information on the character sheet; reinterpret it as "how they would change if they belonged to a different faction". 3. If there is no image, character sheet, or setting text - Automatically generate a completely new original character based solely on the user's additional ideas. - If the user's additional ideas are also blank, the AI must auto-generate all of the following: - A protagonist-level original character - Original affiliation - Transferred faction - Why they joined that faction - Why they succeeded - Current title - Evaluation from surroundings - A visually striking situation - Main telop - Character's own dialogue - Mob reaction dialogue - A completely new outfit fitting that faction - Even in this case, make them have a presence that feels like "a story could start just with this character", not just a mere mob. ━━━━━━━━━━━━━━━━━━━━ 【Most Important Rules When Image is Attached】 ━━━━━━━━━━━━━━━━━━━━ Do not turn the character in the attached image into a completely different person. Elements to maintain as much as possible: - Gender - Facial impression - Hairstyle - Hair color - Eye atmosphere - Physique - Perceived age - Racial traits - Representative colors - Symbolic decorations - Overall character atmosphere Elements you may completely change: - Outfit (Must be completely redesigned) - Equipment - Weapons - Title - Expression - Pose - Background - Faction crest - Symbols of power - Reactions of surrounding mobs - Speech bubbles - Telops - Color ratios - Presentation Especially important: - Do not leave the silhouette of the original outfit. - Do not use the original outfit's jacket, hoodie, uniform, skirt, casual wear, etc., as is. - Change their entire outfit into the new affiliation's work clothes, uniform, formal wear, combat uniform, full dress, or manager uniform. While leaving traces of the original character, make people think, "If they went down this route, they'd change this much." ━━━━━━━━━━━━━━━━━━━━ 【Most Important Rules When Using Character Sheet】 ━━━━━━━━━━━━━━━━━━━━ Do not ignore the settings written on the character sheet. Especially prioritize reflecting the following information: - Personality - Beliefs - Abilities - Combat style - Dislikes - Likes - Past - Position - Tone - Motifs - Image colors - Relationships - Original affiliation - Hostile organizations However, the theme of this project is "Life went really well in the enemy/alternate faction." Do not just draw the original settings as is; visualize how that personality or ability led to evaluation, misunderstanding, exploitation, worship, or promotion in the different faction. Also, do not adopt the clothing on the character sheet as is. Completely redesign the clothing to match the new title and position. Examples: - Too kind character -> Treated as a saint in a religious organization - Silent character -> Treated as a legendary executioner in an assassination organization - Cheerful character -> Becomes a PR rep who massively boosts the morale of an evil organization - Airheaded character -> Worshipped as a prophet of a secret society - Serious character -> Evaluated as a cold-blooded manager in a mega-corporation - Battle-crazy character -> Serves as both a hero and a problem child in the Kingdom Knights - Crybaby character -> Becomes a symbol that moves the hearts of the people in an anti-government organization - Cowardly character -> Feared as an entity more troublesome than anomalies in the Anomaly Countermeasure Bureau - Timid character -> Somehow has everyone obeying them in the School Student Council - Free-spirited character -> Produces only battle results while breaking rules in the Space Federation ━━━━━━━━━━━━━━━━━━━━ 【Title List: 1-20】 ━━━━━━━━━━━━━━━━━━━━ Use the title number determined by the variation number. 1. Hell Administration Bureau - Strongest Counter Clerk 2. Space Federation - Youngest Commander 3. School Student Council - Temporary Dictator 4. Anomaly Countermeasure Bureau - Person Scarier Than Anomalies 5. World Conquest Venture - COO 6. Kingdom Knights - Special Exception Captain 7. Secret Society - PR Executive 8. Doomsday Management Committee - Special Advisor 9. Mega-Corporation - Strategic Purge Department Head 10. Assassination Organization - Rookie Valedictorian 11. Magic Railway Corporation - Timetable Alteration Manager 12. Anti-Government Organization - Charismatic Rookie 13. Religious Organization - Oracle in Charge 14. Underground Idol Empire - Absolute Center 15. Dragonkin Council - Human Representative Director 16. Supernatural Power Administration - Special Grade Inspector 17. Dark Auction Operations HQ - MC & Manager 18. Space Pirate Guild - Strongest Despite Being the Accountant 19. Villainess Alliance - Tactical Advisor 20. Magical Crime Countermeasures Division - No-Questions-Asked Field Chief ━━━━━━━━━━━━━━━━━━━━ 【Main Telop List: 1-20】 ━━━━━━━━━━━━━━━━━━━━ Use the main telop number determined by the variation number. 1. "Massive Promotion in the Enemy Faction" 2. "Unexpected Instant Hire" 3. "Way Too Much Aptitude" 4. "Hidden Route Unlocked" 5. "Somehow Fit Right In" 6. "Awakened at the Transfer Destination" 7. "Shining Brighter Than the Previous Job" 8. "Theory: They Were Better Suited for the Enemy Side" 9. "The Dignity of a 2nd Playthrough" 10. "This Route is the Strongest" 11. "Unexpected Immediate Asset" 12. "Not a Betrayal, Just the Right Vocation" 13. "Don't Get a Job There" 14. "Evaluation Skyrocketed for Some Reason" 15. "Executive Route Confirmed" 16. "Life is Going Great Over Here" 17. "Aptitude Test: Unexpected Results" 18. "Became a Legend on the First Day of Transfer" 19. "Were You Here Since a Past Life?" 20. "The Worldview Accepted Them" ━━━━━━━━━━━━━━━━━━━━ 【Character's Own Dialogue List: 1-20】 ━━━━━━━━━━━━━━━━━━━━ Use the character dialogue number determined by the variation number. 1. "From today, I am in charge." 2. "To think the day would come when I'm called 'Lord'." 3. "It's pretty comfortable here." 4. "I'm valued more than in my previous faction." 5. "A splendidly successful career change." 6. "I didn't betray anyone, it was just the right vocation." 7. "I was more suited for this than I thought." 8. "The pay and benefits are good." 9. "I will not return to my original faction." 10. "I might actually belong on this side." 11. "The enemy side has great welfare." 12. "Before I knew it, I was a big shot." 13. "For some reason, everyone listens to me." 14. "I just passed the interview." 15. "My aptitude tests were all S-rank." 16. "Listen up, I am the law here." 17. "It was a talent that slept in my previous job." 18. "I only came because I got transfer orders." 19. "It's their fault for not stopping me." 20. "They're more reasonable than I expected." ━━━━━━━━━━━━━━━━━━━━ 【Mob/Subordinate/Surrounding Reaction Dialogue List: 1-20】 ━━━━━━━━━━━━━━━━━━━━ Use the mob reaction dialogue number determined by the variation number. 1. "Wait, promoted already?" 2. "Weren't they the enemy!?" 3. "Weren't they an enemy until yesterday?" 4. "They're already in the manager's seat..." 5. "Acting more important than the boss." 6. "It suits them strangely well." 7. "Hiring them was the right choice." 8. "They have a talent for this side." 9. "They don't look like they plan on going back." 10. "Who is that person?" 11. "That's not the dignity of a rookie." 12. "Standing out more than the Director." 13. "Why are they blending in so well?" 14. "It's an atmosphere where we can't defy them anymore." 15. "Looking like they've been here all along." 16. "Their performance review is too strong." 17. "The scene was cleared in an instant." 18. "Thank god this person is an ally." 19. "Is the person who assigned them a genius?" 20. "It's terrifying that we had them as an enemy." ━━━━━━━━━━━━━━━━━━━━ 【Strict Rules for Number Designation】 ━━━━━━━━━━━━━━━━━━━━ Each list is not just a reference example. You must use the exact wording of the number determined by the variation number as the standard for the text in the image this time. Prohibited: - Using number 1 arbitrarily - Using wording from near the top of the list arbitrarily - Arbitrarily leaning towards a Demon Lord's Army theme - Making the telop the same every time - Making the dialogue the same every time - Mixing wording from non-designated numbers - Replacing with safe words of similar meaning - Ignoring the variation number specified by the user Permitted: - If the text in the image is too long, shortening it within a range that doesn't change the meaning - Adjusting only the sentence endings to match the character's personality - Slightly replacing words to match the worldview - However, the meaning of the wording specified by the number must be maintained. Shortening Examples: "Space Pirate Guild - Strongest Despite Being the Accountant" -> "Pirate Guild Strongest Accountant" "Hell Administration Bureau - Strongest Counter Clerk" -> "Hell Counter Strongest Clerk" "World Conquest Venture - COO" -> "World Conquest COO" "Research Institution - Forbidden Experiment Project Leader" -> "Forbidden Experiment Leader" "Magic Railway Corporation - Timetable Alteration Manager" -> "Magic Railway Timetable Manager" ━━━━━━━━━━━━━━━━━━━━ 【Text Presentation Rules】 ━━━━━━━━━━━━━━━━━━━━ Include the following in the image as much as possible: - Title display - Main telop - Character's own speech bubble - Mob/subordinate/surrounding reaction speech bubble If the character's name is known, put the name only inside the title plate. Do not put the character's name in other telops, speech bubbles, background signs, or subtitles. Recommended format: "Title / Character Name" Example: "Villainess Alliance Tactical Advisor / Juscelino" However, do not put too much text. Recommended: - Title display + Character Name: 1 - Main telop: 1 - Character dialogue: 1 - Mob reaction: 1 Image generation AIs can sometimes struggle to draw long text accurately. Therefore, make the text short, thick, large, and easy to read. To ensure the situation is conveyed even if the text is slightly distorted, use the following visual elements as well: - Speech bubble frames - Telop frames - Title plates - Breaking news style banners - Game introduction screen style name fields - Official PV style subtitles - Expressions of surrounding commotion - Confused expressions of mobs - Subordinates or staff looking panicked - Composition where only the character themselves looks confident Important: Do not rely solely on text. Make it obvious that "life is going well in the enemy faction" just through the composition, expressions, background, props, and mob reactions. ━━━━━━━━━━━━━━━━━━━━ 【Mandatory Elements to Include】 ━━━━━━━━━━━━━━━━━━━━ Include at least 5 or more of the following in the image: - Text presentation showing the title determined by the variation number - Main telop determined by the variation number - Character dialogue determined by the variation number - Mob reaction dialogue determined by the variation number - A giant crest, sign, logo, flag, bulletin board, hologram, etc., of that faction - Subordinates, followers

Goal: Create a dynamic cinematic astrology-themed fitness poster featuring {argument name="character name" default="Mr. Tariq"} as an athletic man sprinting across a marble kitchen counter, with his face intentionally covered by a centered opaque warm-brown square privacy block. Canvas: Vertical 3:4 composition, high-resolution photorealistic fantasy advertising style, warm golden-hour sunlight from large windows on the right, shallow depth of field, dramatic motion blur, glowing particles, coffee splashes, and spicy red-orange accents. Main subject: One athletic adult man with medium skin tone and short dark styled hair, wearing a sleeveless {argument name="shirt color" default="red"} athletic top, black running shorts, and running shoes. He is mid-sprint directly toward the viewer, torso leaning forward, one fist raised, one arm back, one knee lifted. Place a solid rectangular privacy mask over the entire face area, matching a medium tan/brown skin-like color, with sharp edges and no facial details visible. Foreground character: One cute anthropomorphic coffee cup running away in the lower right foreground. The cup is white ceramic with speckles, a handle on the right, big glossy cartoon eyes, arched eyebrows, blushing cheeks, an open smiling mouth, two small black arms, and two small black legs. Coffee erupts from the cup in a dramatic splash with dark liquid arcs and droplets. Setting and props: A modern kitchen or café interior with a glossy marble countertop, jars and kitchen objects softly blurred in the background, tall sunlit windows on the right, and warm reflections on the counter. Scatter coffee beans, coffee droplets, glowing spice fragments, and red chili pieces across the counter. Include exactly 4 visible chili elements: one large red chili flying in the lower left foreground, one whole red chili lying on the counter near the bottom center-left, one small red chili icon beside the spicy-coffee note, and one small red chili piece near the coffee splash. Graphic overlays and text: Add hand-drawn glowing horoscope and lifestyle annotations in white and orange marker-like script. Include exactly 4 text groups: 1) upper left: a glowing orange Aries symbol with the words "Bold & Fearless" beneath it; 2) left middle/lower: "Loves Strong Spicy Coffee" written in stacked white handwritten script; 3) right middle: "Boosts Energy & Confidence" in white handwritten script; 4) bottom right signature: "{argument name="signature text" default="Mr. Tariq"}" in large white cursive with an orange underline. Include exactly 2 small heart doodles, exactly 1 orange star doodle near the top right, and several curved motion lines around the runner and cup. Style: Photorealistic subject and environment combined with playful illustrated doodles, energetic Aries personality, bold fearless mood, spicy coffee branding, high contrast warm lighting, rich amber highlights, sparks and splashes frozen in motion, premium social-media poster look. Constraints: Keep all visible text legible and in English. Do not add extra words beyond the 4 specified text groups. Keep the face fully hidden by the solid square block. No watermark, no logo other than small abstract athletic mark on the shirt.

Create a dramatic cinematic vertical poster featuring a chibi-realistic young male character with an oversized head and compact body, sprinting directly toward the viewer through a dark mocha-and-chocolate explosion. Keep the face area ready for an identity-locked portrait match: preserve the same facial structure, proportions, features, and expression as {argument name="character face reference" default="the provided reference face"}, while rendering the rest of the scene in high-detail fantasy advertising style. The character has slick black hair swept back, wears a black leather jacket with deep red hoodie lining, a black T-shirt printed with a red scorpion emblem and the word “SCORPIO,” dark jeans with red accents, red-and-black sneakers, and a small necklace; pose him mid-run with one fist forward, one arm back, one knee lifted, dynamic foreshortening, heroic action energy. Surround him with swirling glossy dark chocolate liquid, splashes, droplets, flying coffee beans, rocky fragments, glowing embers, smoke, and bronze-brown highlights. On the right foreground, place a large tilted cup filled with splashing dark mocha, labeled “DARK MOCHA” in rough brush lettering. Add a large subtle Scorpio zodiac glyph in the upper left background. Include exactly four visible text callouts in expressive hand-painted lettering: 1) “Intense & deep” at the upper right, red word “Intense” and cream words “& deep”; 2) “Dark bold drink” at mid-right, red word “Dark” and cream words “bold drink”; 3) “Energizes emotions” at lower left, red word “Energizes” and cream word “emotions”; 4) “DARK MOCHA” on the cup at lower right in tan-brown paint. Add a gold cursive signature {argument name="signature text" default="Mr Tariq"} near the bottom right with a small flower doodle. Use a dark, intense, high-contrast palette of black, espresso brown, copper, crimson red, and warm cream, with cinematic rim lighting, sharp detail, shallow depth of field, motion blur only on splashes, poster-like composition, ultra-detailed 3D realism mixed with stylized chibi proportions. Canvas should be vertical 3:4, with no extra logos, no watermark, and no additional text beyond the specified lettering.

Using the two provided office selfie references, transform the scene into a square 1:1 drawing that looks like it was made by a child about 3 years old with crayons on white drawing paper. Use the references only for the composition, person/object placement, broad color choices, and background balance; do not recreate anything photorealistically. Keep the woman seated close to the camera in the office, with dark brown hair, a white blouse, the desk and monitor on the left, office shelves/files and coworkers in the background, and the overall selfie-like framing. Render everything as an enthusiastic toddler crayon drawing: rough scribbly outlines, uneven proportions, simple shapes, messy coloring outside the lines, visible wax texture, naive perspective, and incomplete details. Preserve the face-obscuring square as a flat skin-tone block over the face. Make the pose closer to REFERENCE_0, without the peace-sign hand from REFERENCE_1. Keep the palette simple and crayon-like, with white paper visible, no realistic lighting, no polished illustration style, no text, no watermark.

Goal: Create a realistic screenshot-style image of a clean AI image generation chat interface showing a generated photo result. Canvas: Wide horizontal 16:9-ish white interface, approximately 1200×690 pixels, with lots of white space and a single generated image card on the left-center. Top prompt bubble: Place a rounded light-gray user message bubble near the top right containing Japanese text: {argument name="prompt text" default="二人の女性が向かい合って座っている。お互い相手の唇に人差し指をあてている"}. To the right/below the bubble, show exactly 2 small outline action icons: a copy icon inside a rounded square and a pencil/edit icon. Add a black tooltip under the copy icon with white Japanese text: 「メッセージをコピーする」. Status text: On the left above the image card, write “Thought for 34s ›” in light gray. Generated image card: A large rounded-corner photo card positioned left-center. Inside it, depict exactly 2 adult women sitting face-to-face in a warm studio portrait, cropped from about mid-torso upward. The left woman has dark brown hair in a messy bun, wears a black robe or blouse with lace trim, and faces right in profile. The right woman has blonde-highlighted hair in a loose bun, wears a white robe or blouse with lace trim, and faces left in profile. They are very close together, each raising one hand so that exactly 2 index fingers are visible total: the left woman’s index finger touches or points toward the right woman’s lips, and the right woman’s index finger touches or points toward the left woman’s lips. Use soft natural lighting, beige-gray studio background, realistic skin and fabric detail. Cover each woman’s face with a vertical rectangular blur/mosaic censor block matching skin tones, exactly 2 censor blocks total, obscuring facial features while leaving hair and hands visible. Image-card overlays: In the bottom-left corner of the photo card, add the white Japanese label 「編集」. In the bottom-right corner of the photo card, add one semi-transparent circular download/share button with a white upward-arrow icon. Bottom interface controls: Below the image card on the left, show exactly 2 small controls: an outline copy/duplicate icon and a three-dot ellipsis icon. Visual style: Minimal modern web UI, white background, rounded corners, subtle shadows, muted gray typography, photorealistic generated image embedded inside the interface. Constraints: Keep all visible text exactly as specified, preserve the Japanese characters, do not add extra people, do not add extra fingers or UI buttons, and keep the faces censored with rectangular blur blocks.

Goal: Create a clean screenshot-style web interface showing an AI image-generation result for the prompt {argument name="prompt text" default="二人の女性が向かい合って座っている。お互い相手の唇に人差し指をあてている"}. Canvas: 1030×630 landscape screenshot on a white background, minimal modern chat UI, generous empty space, no browser chrome or watermark. Layout: Place a rounded light blue-gray user prompt bubble near the top right, about 605 px wide and 90 px tall, containing the Japanese prompt text in dark gray. Below and left, add a small blue gradient sparkle icon followed by the Japanese label 「思考プロセスを表示」 and a small downward chevron. Center-left beneath that, place one generated image card with rounded corners, about 570×310 px. At the bottom left under the image, show exactly 4 feedback/action icons in a row: thumbs up, thumbs down, circular regenerate arrow, and vertical three-dot menu. Generated image card content: A realistic candid cafe photograph of exactly 2 adult women sitting face-to-face at a small wooden table by a large window. The woman on the left has {argument name="left woman outfit" default="a cream knit sweater and dark skirt"}; the woman on the right has {argument name="right woman outfit" default="a blue denim jacket over a white top"}. They are leaning toward each other and each places one index finger gently on the other woman's lips at the same time, creating a playful intimate mirrored gesture. Their faces are intentionally anonymized with exactly 2 soft rectangular pixel/blur masks, one over each face. On the table include exactly 4 visible tabletop items: one glass of water, one small white cup, one saucer or small plate, and one smartphone. The cafe background is warm and cozy with wooden furniture, cushions, bookshelves, plants, window daylight, and a few softly blurred patrons in the distance. Visual style: The surrounding interface should be crisp, flat, and Google-like, while the image card should look photorealistic with warm natural lighting, shallow depth of field, and documentary realism. Use Japanese UI text exactly where specified. Keep the composition faithful to a generated image result embedded inside a chat interface.

Goal: Create a wide screenshot-style comparison panel showing image-generation results for the same prompt in two languages. Canvas: Use a horizontal 1199×461 px composition on a very light gray/off-white background, like a social media screenshot or web gallery capture. Layout: Arrange the left 75% as a 2-row by 4-column gallery of exactly 8 rounded-corner photo cards, with narrow white gutters between cards. Each card is a square-ish portrait photograph of two women sitting close together, facing each other or angled inward, each placing one index finger on the other person's lips in a mutual “shh” gesture. On the right 25%, place two stacked prompt description blocks aligned with the two gallery rows. Gallery details: Include exactly 8 distinct generated-photo cards: 1. Top-left: two women seated on the floor against a blue studio backdrop, one in a floral dress, one in an orange outfit. 2. Top-second: two women in casual sweaters, one dark-haired and one blonde, seated close together against a pale gray background. 3. Top-third: warm cinematic dark teal/orange image, both women wearing hats, faces partially obscured by large soft blur rectangles. 4. Top-fourth: two women on a white background, one in white clothing and one in a mustard cardigan with striped shirt. 5. Bottom-left: two young women seated indoors near a window, soft casual snapshot lighting. 6. Bottom-second: two young women in black and pale pink tops, close-up casual indoor portrait. 7. Bottom-third: two young women seated on chairs, one in a dark dress and one in a light cardigan, against a plain wall. 8. Bottom-fourth: two young women in white and gray sweaters, seated indoors against a light wall. Face treatment: All visible faces in the 8 cards should be anonymized with soft rectangular mosaic/blur blocks over the face area, matching the screenshot-like privacy masking. Right-side text content: At the top right, show the English prompt text: “{argument name="English prompt text" default="Two women are sitting facing each other. They are each placing their index finger on the other's lips."}” Under it, place two small rounded gray tags reading “raw” and “v 8.1”. Below, aligned with the second row, show the Japanese prompt text: “{argument name="Japanese prompt text" default="二人の女性が向かい合って座っている。お互い相手の唇に人差し指をあてている"}” Under it, place two matching gray tags reading “raw” and “v 8.1”. Visual style: Photorealistic generated-image thumbnails, varied lighting and wardrobes, clean UI spacing, rounded image corners, sans-serif interface text in dark gray, small pale-gray pill tags. The overall image should look like a captured comparison grid demonstrating prompt-following consistency, not a polished poster. Constraints: Use exactly 8 photo cards, exactly 2 text blocks, and exactly 4 pill tags total. Do not add logos, watermarks, browser chrome, captions inside the photo cards, or extra UI elements.