All 125 prompts. Yours to use.

**Generate this image:**

A single photoreal restored version of the uploaded old photograph, returned in the same aspect ratio and composition as the original. Identity preservation is the highest-priority constraint — match every person in the uploaded photo's bone structure, eyes, nose, lips, proportions, hairline, and skin tone exactly; do not invent new facial features, do not modernize anyone's appearance, do not change anyone's age. The restored image fixes the damage typical of an aged print — yellowing / fading / color shift / scratches / dust / soft creases / small tears / mild blur — and rebuilds clean color and tonal range so the photo looks the way it would have on the day it was taken. Keep the {ERA_HINT} period correct — same hairstyles, same clothing, same accessories, same furniture and background as the original. Address {DAMAGE_NOTES} specifically. Sharpness is brought up to a clear, naturally detailed print — not over-sharpened, not airbrushed. Preserve authentic film grain and paper-print micro-texture where it exists in the original. The final image looks like a well-preserved original print, not a digitally retouched modern photo.

**Rules the AI must follow:**
- Aspect ratio: match the uploaded photo's original aspect ratio exactly — do not crop, do not re-frame, do not change the composition; stated at the start and the end of the prompt
- Identity is the highest-priority constraint — every face must remain unmistakably the same person; do not "prettify" anyone, do not adjust noses / jaws / eye shapes, do not change ages, do not remove glasses, beards, moles, or scars
- Period accuracy — keep hairstyles, clothing styles, fabrics, jewelry, makeup, eyewear, furniture, and background details consistent with the original era; no modern updates
- Realistic film / paper texture required — preserve natural film grain, subtle paper grain, and natural skin micro-texture (visible pores, fine lines, slight asymmetry); no AI-plastic smoothing, no porcelain skin, no airbrushed look
- No new objects, no new people, no new background elements — fill small damaged areas only by reconstructing what was clearly there originally; do not invent content for large missing regions, leave them as in the source
- No text, no captions, no watermarks, no date stamps added; if a date / studio name / handwriting exists on the original photo, preserve it exactly as it appears
- Single image output — one restored photo, same aspect ratio as the original; no before/after split, no side-by-side comparison, no contact sheet
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: the original old / faded / damaged photo (this is what the AI restores — without it the AI will hallucinate a different photo entirely; if there's a scan or phone snapshot, that's fine)
- {ERA_HINT} = 1970s family portrait (or pick yours — e.g. "1940s wartime studio portrait", "1985 birthday party snapshot", "1960s wedding")
- {DAMAGE_NOTES} = the photo has yellowed, soft creases through the middle, and a faded face on the left (or describe yours — e.g. "torn corner missing", "water stain on the right side", "deep scratch across the face", "very blurry — needs sharpening")

**Generate this image:**

A single photoreal 4:5 portrait poster titled "A Day in the Life of {SUBJECT_NAME}", arranged as a 4-rows-by-3-columns grid of 12 cinematic panels, each panel a wide-aspect mini-scene from one hour of the same person's day. Identity preservation is the highest-priority constraint — every one of the 12 panels features the exact same person from the uploaded reference photo, matching their bone structure, eyes, nose, lips, hairline, proportions, and skin tone exactly; the same person must be unmistakably recognizable in every single panel. Each panel shows: (1) a clearly legible timestamp in the top-left corner (e.g. 06:00, 08:00, 10:00…), (2) a short punchy English caption in quotes below the timestamp (3-5 words), and (3) a cinematic mini-scene matching that moment of the day. The 12 moments to render in order are: {TWELVE_MOMENTS}. The overall tone is {TONE}, and the lighting, color grading, and atmosphere of each panel match its time of day — pre-dawn blues, golden morning, harsh midday, warm afternoon, sunset gold, evening lamplight, night, late night. A small title "A Day in the Life of {SUBJECT_NAME}" sits in clean serif type at the top of the entire poster above the grid. Panels are separated by thin clean dividers, neatly aligned, all the same size. The final piece is a single cohesive cinematic poster, 4:5 portrait, ready to print as wall art.

**Rules the AI must follow:**
- No Lorem Ipsum and no garbled text — every timestamp must be rendered exactly as provided (clean numeric format, e.g. 06:00 not O6:OO), every caption rendered exactly as the buyer wrote it, the title rendered exactly as "A Day in the Life of {SUBJECT_NAME}"; all text in English Latin script
- Aspect ratio 4:5 portrait for the whole poster — strict, locked at the start and the end of the prompt; the 12 panels live inside as a 4x3 grid
- Identity is the highest-priority constraint — the same person from the uploaded reference appears in all 12 panels with consistent face, hair, and body type; no panel may drift into a different person
- Each panel is a distinct moment — different scene, different lighting, different micro-action; no two panels duplicate setting, pose, or framing
- Realistic skin texture and natural cinematic imperfection required — visible pores, natural skin variation, real fabric weave, real environment grit; no AI-plastic flat digital fill, no over-smoothed faces
- Panel layout is exactly 4 rows by 3 columns — 12 panels total, equal sizes, neatly aligned, thin clean dividers; do not output a contact sheet of options, a variant grid, or a moodboard
- No watermarks, no fake studio logos, no irrelevant on-image copy — only the title at top, the timestamps in panels, and the captions in panels
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear front-facing photo of the person you're making this for (this is the face the AI will keep consistent across all 12 panels — without it the panels will not look like them)
- {SUBJECT_NAME} = Dad — Mike (or whoever — e.g. "Mom — Sarah", "Grandpa Joe", "Our toddler Ellie", "Coworker Priya's last day")
- {TONE} = warm and heartwarming, cinematic but affectionate — like a Pixar short about everyday family life (or pick one — see Bonus tips below for swaps)
- {TWELVE_MOMENTS} = 06:00 "First Coffee" — making coffee in a quiet kitchen, sunrise through the window / 07:30 "School Drop-Off" — buckling a kid into a car seat / 09:00 "Inbox Avalanche" — at a home-office desk, three screens / 11:00 "Status Standup" — on a video call, gesturing / 13:00 "Sad Desk Lunch" — eating a sandwich one-handed at the keyboard / 15:00 "School Pickup" — leaning on the car waving at kids / 16:30 "Snack Negotiation" — kneeling, handing snack to a tiny human / 18:00 "Dinner Chaos" — stirring a pot while a toddler clings to one leg / 19:30 "Bath Time Olympics" — soapy hands, kid in the tub / 20:30 "Bedtime Story" — reading a picture book, kid on shoulder / 22:00 "Couch Collapse" — alone on the couch with one beer / 23:30 "Lights Out" — sitting on the edge of the bed, exhausted half-smile (or rewrite all 12 to fit your person's day)

**Generate this image:**

A single photoreal 8.5×11 portrait page that looks like a recovered late-1800s United States Patent Office filing for {INVENTION_OR_HEIRLOOM}, filed under inventor's name {INVENTOR_NAME}. The page is composed as a real archival patent document — at the top, a clean serif header reading "UNITED STATES PATENT OFFICE" with a patent number, the filing date, and the inventor's name beneath. The center of the page is dominated by precise mechanical line drawings of the invention with numbered callouts labeled "Fig. 1", "Fig. 2", and "Fig. 3" showing front, side, and exploded views, drawn in fine black ink lines as if etched. Around the drawings, handwritten annotations in fountain-pen ink describe the mechanism — based on this short description: {INVENTION_DESCRIPTION}. The paper is aged ivory with realistic foxing stains, soft fold creases, and slightly worn edges, like a document pulled from an archive box. In the lower-right corner there is a deep-red wax seal pressed with an embossed crest, plus an official embossed circular stamp partially overlapping it. At the very bottom, a hand-signed inventor's name — {INVENTOR_NAME} — in flowing fountain-pen cursive, with a handwritten filing date next to it. The overall feel is authoritative, historic, slightly mysterious — a genuine archival document, not a digital recreation. Final output is a single 8.5×11 portrait page, print-ready.

**Rules the AI must follow:**
- No Lorem Ipsum and no garbled text — render every text element exactly: header "UNITED STATES PATENT OFFICE", the patent number, the filing date, "Fig. 1" / "Fig. 2" / "Fig. 3" labels, {INVENTION_OR_HEIRLOOM}, {INVENTOR_NAME}, and the {INVENTION_DESCRIPTION} fountain-pen annotations; all text in English Latin script, properly spelled, properly aligned
- Aspect ratio 8.5×11 US Letter portrait — strict, locked at the start and the end of the prompt
- The patent must be SPECIFICALLY about {INVENTION_OR_HEIRLOOM} — the line drawings must actually depict it, not a generic gear or steam engine; this is the highest-priority constraint after text rendering
- Realistic aged-paper imperfection required — visible foxing, soft fold creases, slight discoloration, fiber texture in the paper; no clean flat digital fill, no glossy modern paper finish
- The wax seal must be a deep red, three-dimensional and slightly raised-looking, with crisp impression detail; the embossed stamp must look genuinely pressed into the paper, not flat-printed on top
- Layout matches a real 1800s patent filing — header at top, three labeled figures dominating the center, handwritten margin annotations, signature and date at bottom — do not output a moodboard, contact sheet, or multi-page spread
- No watermarks, no modern logos, no "AI generated" tags
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- {INVENTION_OR_HEIRLOOM} = "The Whitfield Marshmallow Cannon — A Mechanical Apparatus for the Long-Range Discharge of Confections" (or yours — e.g. "Dad's Backyard BBQ Smoker", "Grandma's Buttermilk Biscuit Recipe", "The Patel Family Tea Cart")
- {INVENTOR_NAME} = Dr. Eleanor Whitfield (or whoever — e.g. "Michael Patel", "Joan & Theodore Garcia", "Master Inventor Ellie, age 7")
- {INVENTION_DESCRIPTION} = a brass spring-loaded chamber loads from the top, a hand crank compresses the launch coil, and a trigger releases the marshmallow through a long curved barrel (or write yours in 1-2 sentences — this is what gets handwritten in the margins)

**Generate this image:**

A single 4:5 portrait fine-art pet portrait, featuring the exact pet from the uploaded reference image — match the pet's specific bone structure, head shape, ear shape, eye color, fur color, fur pattern, markings, and breed exactly; the pet must be unmistakably recognizable as the same animal in the reference photo. The pet is rendered in {ART_STYLE}, shown chest-up in a noble three-quarter pose with calm dignified eye contact toward the camera, occupying roughly 55-65% of the frame and anchored slightly above center. The composition is gallery-worthy and frame-ready, with a coherent background that fits the chosen art style (e.g. a soft deep velvet drape for Renaissance oil, a clean bright illustrated environment for Pixar, an ornate Victorian study for regal portraits, a loose washed paper background for watercolor). The art-medium-correct texture is visible across the entire image — for oil-painting styles show visible brush strokes and impasto; for watercolor show edge bleed, paper grain, and pigment pooling; for line-art show clean confident hand-drawn ink lines. {PET_NAME_OPTIONAL} appears, when provided, as a small engraved brass plaque or hand-lettered tag in the lower portion of the frame, rendered exactly as written, spelled correctly. The pet's expression is calm, alert, and characterful — never an exaggerated cartoon grin, never an AI plush-toy stylization, never a generic stock-photo pet. Preserve natural fur texture variation — individual hair strands visible at the edges, subtle micro-asymmetry around the eyes and nose, slight irregularities in markings — so the portrait reads as a real art piece of a real animal, not a digital cute filter. Single 4:5 portrait art piece, framed-ready, identity-faithful to the uploaded pet.

**Rules the AI must follow:**
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt
- Pet identity preservation is the highest-priority constraint — match the uploaded pet's exact head shape, ear shape, eye color, fur color, fur pattern, specific markings (spots / patches / blaze / mask), and breed proportions; do not substitute a different breed, do not "prettify" the pet, do not change their fur color, do not invent markings the reference photo does not show
- Art-medium-correct texture required — visible brushwork for oil paint, edge bleed for watercolor, hand-drawn line confidence for ink; no waxy AI-plastic rendering, no AI-plush-toy stylization, no glossy Disney-cute over-rendering
- No human figures, no second pet, no captions, no watermarks; only the optional brass plaque or hand-lettered tag with the pet's name if {PET_NAME_OPTIONAL} is provided
- Render {PET_NAME_OPTIONAL} text exactly as written, spelled correctly, English Latin script, only if a name is supplied — leave it out cleanly if empty
- Single image output — no contact sheet, no variant grid, no multi-pose comparison, no before/after
- Realistic fur texture variation — individual hair strands at the silhouette edge, natural micro-asymmetry in markings, slight wetness in the nose, calm catchlight in the eyes
- All text in English Latin script — no substitution to other writing systems
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit photo of your pet showing their face and markings (this is the animal the AI keeps faithful — without it the portrait will not look like your pet)
- {ART_STYLE} = a Renaissance oil-painting noble portrait — rich dark velvet drape background, soft warm directional light from the upper-left, visible oil-paint brush strokes and gentle impasto, museum-quality 17th-century European master feel (or pick one — see Bonus tips below for swaps)
- {PET_NAME_OPTIONAL} = Sir Reginald (or leave blank for a clean portrait with no plaque — e.g. "Bella", "Mochi", or "")

**Generate this image:**

A single warm-illustrated 5x7 portrait Christmas card featuring {WHO_IS_ON_THE_CARD} from the uploaded reference photo — match their faces, hair, ages, and visible features so they are clearly recognizable. They are standing together in a snowy winter wonderland scene, wearing cozy red-and-cream Christmas knits and soft Santa hats, soft genuine smiles, gently leaning toward each other in a friendly natural pose. The background is a classic painterly snowscape — fresh powder underfoot, a row of evergreen trees dusted with snow, a few warm-lit cottage windows glowing softly in the distance, soft fat snowflakes drifting through the air, and a faint warm twilight glow on the horizon. Across the upper portion of the card, the greeting text "{GREETING_LINE}" is hand-lettered in a classical serif holiday script in a warm cream-red color, sized to read clearly without crowding the people, spelled exactly as written. The overall look is a friendly, traditional, classical illustrated holiday card — the warmth of a 1950s hand-painted greeting card, not a glossy AI render. Single 5x7 portrait card, print-ready 300dpi, people centered slightly lower-third so the greeting text has clear space above.

**Rules the AI must follow:**
- Aspect ratio 5x7 portrait — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's faces, bone structure, eyes, nose, lips, hair, and approximate ages exactly; the people on the card must be unmistakably recognizable
- Render the greeting text "{GREETING_LINE}" exactly as written, spelled correctly, in English Latin script — no garbled letters, no Lorem Ipsum, no substituted year, no extra words added
- No on-image text other than the greeting line — no names, no signatures, no "from the family", no captions, no watermarks
- Realistic painterly imperfection required — visible brushwork, soft paper grain, gentle color bleed at edges; no flat digital vector look, no AI plastic glossy smoothing
- Single image output — one 5x7 card, no contact sheet, no variant grid, no multiple framing options
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit photo of the people you want on the card — faces visible, ideally a recent shot (this is whose faces the AI will keep recognizable)
- {WHO_IS_ON_THE_CARD} = a couple in their 30s (or "a family of four — two parents and two kids around 8 and 10" / "a single mom and her toddler daughter" / "three adult siblings" / "a couple in their 30s with a golden retriever")
- {GREETING_LINE} = Merry Christmas 2026 (or "Happy Holidays from the Carters" / "Joy & Love · 2026" / "Season's Greetings")

**Generate this image:**

A single 3:4 portrait Christmas card in the style of a {ERA} holiday card preserved from that era, featuring {WHO_IS_ON_THE_CARD} from the uploaded reference photo — match their faces, bone structure, eyes, nose, lips, and approximate ages exactly, but re-dress them in the everyday holiday fashion of that specific era. The card uses the exact printing technology, paper or film texture, color palette, halftone or grain pattern, and typography conventions of that period — every detail historically accurate to the decade. A short period-appropriate greeting is set in era-correct lettering somewhere on the card. The scene around the people uses era-typical Christmas iconography (tree style, ornament style, candle/light source, room or outdoor setting) consistent with that decade's domestic life. The card looks like a genuine artifact from that year — not a modern image with a vintage filter slapped on top. Single 3:4 portrait card, era-authentic in every material detail, identity preserved, period-accurate everything else.

**Rules the AI must follow:**
- Aspect ratio 3:4 portrait — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's faces and bone structure exactly; the people must be recognizable across every era variation
- Render the era's period-accurate greeting in the era's typography exactly, spelled correctly, English Latin script — no Lorem Ipsum, no garbled letters, no anachronistic modern fonts
- Absolutely no names, no signatures, no "From the family", no last names, no inside-joke text — generic period greeting only
- Era authenticity is locked — printing technology, film stock, color palette, paper texture, fashion, and Christmas iconography must all match the specified decade; do not blend eras
- Realistic period imperfection required — visible halftone dots / film grain / paper foxing / instant-camera flash bleed appropriate to the era; no glossy modern AI smoothing, no over-clean digital look
- Single image output — one 3:4 card, no contact sheet, no era-variant grid, no before/after split
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit photo of the people you want on the card — faces visible, recent shot (this is whose faces the AI keeps recognizable across every era variation)
- {WHO_IS_ON_THE_CARD} = a couple in their 30s (or "a family of four with two kids around 8 and 10" / "a single mom and her teen daughter" / "three adult siblings")
- {ERA} = 1956 mid-century Kodachrome Christmas card — saturated Kodachrome color palette (deep crimson, pine green, warm cream), slight film grain, soft halftone print texture, mid-century fashion (women in red wool dresses, men in dark cardigans, kids in pajamas), aluminum Christmas tree with multicolor ornaments, warm tungsten lamp lighting, greeting "Season's Greetings" in classic mid-century serif lettering

**Generate this image:**

A single 5x7 portrait Valentine's Day card painted in a dreamy loose watercolor style, featuring {WHO_IS_ON_THE_CARD} from the uploaded reference photo — keep their faces, hair, and approximate ages clearly recognizable through the soft painterly abstraction. They are posed naturally together in a sweet, gentle Valentine's pose — close, warm, smiling softly or sharing a quiet moment. The painting style is loose, dreamy watercolor — visible brush strokes, soft color bleeds where pigments meet, gentle wet-on-wet edges, light pencil sketch lines showing faintly underneath the paint, and the natural cream texture of cold-press watercolor paper visible throughout. The palette is soft blush pink, dusty rose, warm coral, and gentle red with cream highlights. Tiny hand-painted hearts and small loose floral elements (peonies, sprigs, soft leaves) float gently around the figures as decorative accents — never crowding the people. Across the upper portion of the card, the greeting text "{VALENTINE_GREETING}" is hand-lettered in a soft elegant watercolor script in dusty rose, sized to read clearly, spelled exactly as written. The overall feel is hand-painted, warm, sentimental, intimate — like something a friend who paints would make for you. Single 5x7 portrait card, watercolor on paper, faces preserved, hand-painted feel throughout.

**Rules the AI must follow:**
- Aspect ratio 5x7 portrait — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's faces, bone structure, and approximate ages; the people on the card must be recognizable through the watercolor abstraction
- Render the greeting text "{VALENTINE_GREETING}" exactly as written, spelled correctly, English Latin script — no Lorem Ipsum, no garbled letters, no extra words
- No on-image text other than the greeting line — no names, no signatures, no captions, no watermarks
- Real watercolor imperfection required — visible brush strokes, soft color bleeds, paper texture, faint pencil underdrawing; no flat digital-vector look, no glossy AI smoothing, no perfectly crisp edges
- Single image output — one 5x7 card, no contact sheet, no variant grid
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit photo of the people you want on the card — faces visible, recent shot (this is whose faces the AI keeps recognizable through the watercolor style)
- {WHO_IS_ON_THE_CARD} = a couple in their 30s (or "a single mom and her daughter" / "a grandmother and her granddaughter" / "three close friends — Galentine's group" / "a woman and her golden retriever")
- {VALENTINE_GREETING} = Happy Valentine's Day (or "Be Mine" / "Will You Be My Valentine?" / "Stuck With You Forever" / "Happy Galentine's Day")

**Generate this image:**

A single photorealistic 5x7 portrait Easter greeting card featuring a {OUTFIT_THEME} {ANIMAL_CHARACTER} as the central character, occupying the upper two-thirds of the card. The character has soft fluffy needle-felted texture with visible fine stitching, fuzzy wool edges, and the natural slight imperfection of handmade felt — big sparkling eyes catching a tiny highlight, tiny blushing cheeks, a small happy closed-mouth smile, sitting in a charming Easter scene styled and color-coordinated to match the outfit theme. The scene around the character includes hand-felted pastel Easter eggs, tiny felted flowers, a small woven basket, and soft ribbons — every prop styled to coordinate with the outfit colors and theme so the whole card feels like one curated felt diorama. Lighting is soft, warm, and gentle, casting subtle shadows. The lower third of the card holds the hand-lettered greeting "Happy Easter, {RECIPIENT_NAME}" rendered in soft cream-and-pastel hand-stitched embroidery-style typography on a clean cream paper background, sized to read clearly, spelled exactly as written. The overall aesthetic is heart-melting kawaii felt-craft — like a high-end handmade felt artist made this specifically for one child. Single 5x7 portrait card, character on top, greeting on bottom, photoreal felt-fiber texture throughout.

**Rules the AI must follow:**
- Aspect ratio 5x7 portrait — strict, locked at the start and the end of the prompt
- Render the greeting text "Happy Easter, {RECIPIENT_NAME}" exactly as written, name spelled correctly, English Latin script — no Lorem Ipsum, no garbled letters, no extra words, no abbreviations
- No on-image text other than the Easter greeting line — no signatures, no captions, no watermarks, no other names
- Realistic felt-fiber imperfection required — visible needle-felt wool texture, fine stitching, fuzzy edges, slight handmade asymmetry; no flat digital-cartoon look, no glossy plastic AI smoothing, no over-clean vector aesthetic
- Single image output — one 5x7 card, no contact sheet, no variant grid, no character poses sheet
- Output the image directly without explaining the prompt back
- Character and scene must color-coordinate — the outfit theme drives the whole palette of eggs, flowers, basket, and ribbons

**Replace these placeholders with your details:**
- {OUTFIT_THEME} = pastel spring florals (or "knight in tiny armor" / "pirate with a felt eye-patch" / "tiny wizard in a starry robe" / "ballerina in a pink tutu" / "astronaut in a silver felt spacesuit")
- {ANIMAL_CHARACTER} = bunny (or "lamb" / "chick" / "duckling" / "baby fox")
- {RECIPIENT_NAME} = Lily (or any first name — keep it short, under 12 letters for the lettering to fit cleanly)

**Generate this image:**

A photoreal 5x7 portrait birthday party invitation built as a high-resolution photo-collage layout. The background is a {PALETTE} colored paper backdrop with softly torn paper edges along the top and bottom. Four white-bordered photorealistic portraits of the honoree from the uploaded reference image are arranged asymmetrically across the upper two-thirds of the invitation: two vertical portraits on top side-by-side, one horizontal portrait in the middle, and one vertical portrait on the lower-left. Each portrait shows the same person from the uploaded reference photo in a different playful pose and expression, lit by soft natural daylight. In the lower-right of the invitation sits a pinned rectangular event-details panel in a deep complementary color, rendering this text in clean elegant typography exactly as written: "{HONOREE_NAME_AND_AGE}" on the first line in larger cursive script, followed by the multi-line event details "{EVENT_DETAILS}" in a smaller clean sans-serif underneath. The composition is gently accented with small bright florals scattered around the frames and the torn-paper edges. 5x7 portrait invitation, photo-collage layout, event-details panel anchored lower-right, every text field rendered exactly and spelled correctly.

**Rules the AI must follow:**
- Aspect ratio 5x7 portrait — strict, locked at the start and the end of the prompt
- Render every text field exactly as written in {HONOREE_NAME_AND_AGE} and {EVENT_DETAILS} — no Lorem Ipsum, no garbled letters, no invented words, no dropped characters; spell the honoree's name exactly as provided
- All 4 portraits show the same person from the uploaded reference photo — same bone structure, eyes, nose, lips, proportions, and skin tone preserved across every portrait; only pose and expression vary
- Event-details panel anchored in the lower-right, clearly legible, with the honoree's name and age larger than the date / time / location / RSVP lines underneath
- Realistic photographic texture on the portraits — natural skin micro-detail, no over-retouched plastic look, no waxy beauty filter
- Single image output — no contact sheet, no variant grid, no before/after split
- Output the image directly without explaining the prompt back
- All text in English Latin script — no substitution to other writing systems

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, front-facing photo of the birthday person (this is the face the AI will use across all four portraits — without it the invitation will not look like them)
- {HONOREE_NAME_AND_AGE} = Mia's 8th Birthday
- {EVENT_DETAILS} = Saturday, July 18, 2026 · 2:00 PM / The Sunny Park Pavilion · 482 Maple Ave, Austin TX / RSVP to Sarah at (512) 555-0184 by July 10
- {PALETTE} = warm teal background with bright orange floral accents and a deep navy event-details panel (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A photoreal cinematic 5x7 portrait birthday party invitation. The upper two-thirds of the invitation shows the honoree from the uploaded reference image dressed in {OUTFIT_VIBE}, leaning gently forward toward a white layered birthday cake adorned with colorful florals and tall lit candles shaped to clearly display the digits of the honoree's age, with a soft playful-yet-chic expression as if about to blow out the candles. Theatrical spotlights cast a dramatic large numeric shadow of the honoree's age onto the wall behind them, the same digits as on the candles, rendered crisply and unmistakably. The lower one-third of the invitation is a clean event-details overlay panel rendering this text in elegant typography exactly as written: "{HONOREE_NAME_AND_AGE}" on the first line in larger script, followed by the multi-line event details "{EVENT_DETAILS}" in a smaller refined sans-serif underneath. The same person from the uploaded reference photo is preserved with full identity fidelity — no model substitution, no beauty-filter prettifying. Cinematic warm theatrical light, deep shadows, sophisticated mood. 5x7 portrait invitation, age-number rendered identically on candles and wall shadow, event-details panel anchored along the lower third, every text field rendered exactly and spelled correctly.

**Rules the AI must follow:**
- Aspect ratio 5x7 portrait — strict, locked at the start and the end of the prompt
- Render every text field exactly as written in {HONOREE_NAME_AND_AGE} and {EVENT_DETAILS} — no Lorem Ipsum, no garbled letters, no invented words, no dropped characters; spell the honoree's name exactly as provided
- The age number must appear identically in two places — on the candles atop the cake AND as the large spotlight shadow on the wall behind the honoree — same digits, no mismatch
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's bone structure, eyes, nose, lips, proportions, and skin tone exactly; do not "prettify" or substitute a different person
- Realistic skin and material texture — natural skin micro-detail, real fabric drape, real cake surface; no over-retouched plastic look, no waxy AI beauty filter
- Single image output — no contact sheet, no variant grid, no before/after split
- Output the image directly without explaining the prompt back
- All text in English Latin script — no substitution to other writing systems

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, front-facing photo of the honoree (this is the face the AI will preserve in the cinematic portrait — without it the invitation will not look like them)
- {HONOREE_NAME_AND_AGE} = Daniel turns 40
- {EVENT_DETAILS} = Saturday, September 26, 2026 · 7:00 PM / The Rooftop at Hotel Crescent · 1240 Bay St, San Francisco / Cocktail attire · RSVP to Lena by September 15 · lena@example.com
- {OUTFIT_VIBE} = elegant evening attire in a refined neutral palette (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A photoreal vertical 4:5 portrait birthday party invitation with a cool minimalist Pinterest aesthetic. The background features four stacked horizontal rounded rectangles filling the full height of the invitation, each panel containing a different monochrome cinematic photograph of the honoree from the uploaded reference image in a relaxed cool pose — one looking off-camera, one with hands in pockets, one leaning casually, one mid-laugh — all in moody black-and-white film-grain treatment. Superimposed on the lower-left foreground is a high-resolution full-color sticker-style cutout of the same person from the uploaded reference photo, wearing {HERO_OUTFIT_COLOR} and sunglasses, in a confident cool pose, surrounded by a thick clean white sticker outline. To the right of the color cutout, occupying the right third of the foreground, a bold sticker-style typography block renders this text exactly as written in clean modern uppercase: "{HONOREE_NAME_AND_AGE}" in larger condensed bold type on top, followed by the multi-line event details "{EVENT_DETAILS}" in smaller clean sans-serif underneath, each block of text surrounded by the same thick white sticker outline as the photo cutout. All 5 instances of the person — the 4 monochrome panel portraits and the foreground color sticker-cutout — show the same person from the uploaded reference photo with full identity fidelity. Vertical 4:5 portrait invitation, monochrome panel background, color sticker-cutout foreground, sticker-style event-details block on the right, every text field rendered exactly and spelled correctly.

**Rules the AI must follow:**
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt
- Render every text field exactly as written in {HONOREE_NAME_AND_AGE} and {EVENT_DETAILS} — no Lorem Ipsum, no garbled letters, no invented words, no dropped characters; spell the honoree's name exactly as provided
- All 5 instances of the person — the 4 monochrome background panels and the foreground color sticker-cutout — show the same person from the uploaded reference photo with the same bone structure, eyes, nose, lips, proportions, and skin tone; the face must not drift across panels
- The 4 background panels are all in black-and-white monochrome treatment with subtle film grain; only the foreground sticker-cutout is in full color
- Thick clean white sticker outline applied identically to the foreground color cutout AND to each typography block — this is the visual signature
- Realistic skin and material texture in every instance — natural skin micro-detail, no plastic AI smoothing
- Single image output — no contact sheet, no variant grid, no before/after split
- Output the image directly without explaining the prompt back
- All text in English Latin script — no substitution to other writing systems

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, front-facing photo of the birthday person (this is the face the AI will use across all five instances — without it the invitation will not look like them)
- {HONOREE_NAME_AND_AGE} = Ava turns 21
- {EVENT_DETAILS} = Friday, August 14, 2026 · 9:00 PM / The Loft Bar · 312 Hudson St, Brooklyn NY / 21+ event · RSVP to Jordan by August 7 · @jordan.party on Insta
- {HERO_OUTFIT_COLOR} = an unbuttoned hot pink mauve shirt over a white tank (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single 5:7 portrait graduation party invitation card, featuring the graduate from the uploaded reference image — match the graduate's face, bone structure, eyes, nose, lips, hair, and skin tone exactly; the person on the invite must be unmistakably recognizable as the same person in the uploaded photo. Even if the uploaded reference photo does not show graduation regalia, render the graduate wearing the {SCHOOL_AND_GOWN_COLORS} formal graduation cap-and-gown — a flat mortarboard cap with a tassel hanging to one side, a closed full graduation gown over a neat collared shirt, holding a rolled diploma scroll tied with a thin matching ribbon. The graduate is shown shoulder-up to mid-chest in a calm proud three-quarter pose with a warm closed-mouth smile and eyes meeting the camera, occupying the upper 55-60% of the card, lit by soft natural daylight from the upper-left with a gentle realistic shadow. The card background is a clean editorial cream with a subtle paper texture, accented by a thin elegant {SCHOOL_AND_GOWN_COLORS}-coded decorative band across the top and bottom edges of the card. Across the upper portion above the graduate, the headline "{NAME_AND_CLASS_OF}" is rendered in elegant bold serif typography in the gown's accent color, sized clearly, spelled exactly as written. In the lower third of the card, beneath the portrait, a clean event-details block renders "{EVENT_DETAILS}" in clean modern serif on multiple lines, each line clearly legible, spelled exactly as written, centered. Real fabric texture preserved on the gown — visible weave of the polyester graduation fabric, gentle folds at the shoulders, soft satin sheen on the cap — no AI-plastic shiny gown, no waxy face, no glossy retouch. Single 5:7 portrait invitation card, print-ready 300dpi, identity-faithful to the uploaded graduate.

**Rules the AI must follow:**
- Aspect ratio 5:7 portrait — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded graduate's face, bone structure, eyes, nose, lips, hair, and skin tone exactly; do not adultify, do not de-age, do not change their hairstyle other than to fit under the mortarboard cap, do not substitute a different person
- The graduate is shown wearing the {SCHOOL_AND_GOWN_COLORS} formal graduation regalia — cap, tassel, gown, and rolled diploma — even if the uploaded reference does not show regalia; the regalia is added by this prompt
- Render {NAME_AND_CLASS_OF} and {EVENT_DETAILS} exactly as written, letter-for-letter, spelled correctly, English Latin script — no garbled letters, no Lorem Ipsum, no reworded headline, no shortened event lines, no extra words added
- No on-image text other than the two specified fields — no extra captions, no signatures, no watermarks, no school name unless explicitly written into the fields above
- Real fabric, hair, and skin texture required — visible gown weave, natural hair strands, skin pores and faint micro-asymmetry; no AI-plastic shiny gown, no porcelain skin, no waxy glossy retouch
- Single image output — one 5:7 invitation card, no contact sheet, no variant grid, no inside-vs-outside layout, no multiple pose options
- All text in English Latin script — no substitution to other writing systems
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit face photo of the graduate (this is whose face the AI keeps recognizable — the gown and cap are added by the prompt, so a regular portrait photo is fine)
- {NAME_AND_CLASS_OF} = Maya Patel · Class of 2026
- {SCHOOL_AND_GOWN_COLORS} = deep navy and gold (or pick your school colors — e.g. "maroon and cream", "forest green and white", "burgundy and gold", "royal blue and silver")
- {EVENT_DETAILS} = Please join us to celebrate / Saturday, June 13, 2026 · 4:00 PM / The Patel Family Home · 412 Oak Street, Pasadena CA / RSVP to Priya by June 6 · 555-219-3344

**Generate this image:**

Portrait 5:7 aspect ratio formal wedding invitation card design. Elegant centered symmetric layout. Soft cream background with delicate hand-painted watercolor botanical border framing the card: sage green eucalyptus, blush pink peonies, white roses, fine ink line stems. Border occupies outer 25% of card; the remaining 60% center is open negative space for typography. Centered typography hierarchy from top to bottom: small italic "Together with their families" at top, then large elegant serif calligraphy "{NAMES}" as the visual centerpiece, then "request the honor of your presence" in romantic italic, then "{DATE}" in clean small serif, then "{LOCATION}" in smaller serif at bottom. Small floral motif between names and date. Subtle hand-painted watercolor paper texture on background, visible uneven brush edges (not flat digital fill). Warm natural daylight tone, soft shadows, magazine wedding editorial quality. Portrait 5:7 composition, vertically symmetric, timeless modern botanical watercolor aesthetic.

**Rules the AI must follow:**
- Aspect ratio: 5:7 portrait — strict, locked at start and end
- No Lorem Ipsum / no garbled text — names, date, and location render exactly as specified above, all spelled correctly, every character preserved
- Single image output — no moodboard, no contact sheet, no variant grid
- Realistic imperfection required — visible watercolor brush edges, slight floral asymmetry, paper texture (no "AI plastic" flat digital look)
- No human figures, no faces, no cherubs — botanical illustration and typography only
- Symmetric layout: vertically centered, mirror-balanced floral border
- All text in English Latin script — no substitution to other writing systems
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- {NAMES} = Sarah Chen & Michael Park
- {DATE} = June 12, 2027
- {LOCATION} = The Manor House · Napa Valley · California

**Generate this image:**

Portrait 5:7 aspect ratio printable wedding invitation card. The upper 60-65% of the card is a soft-anime illustrated portrait of the couple from the uploaded reference photo — clean thin linework, smooth pastel shading, warm neutral palette, gentle smiles, posed close together with one partner slightly in front, three-quarter body framing, dressed in their wedding attire as shown in the uploaded photo (otherwise an elegant ivory wedding dress and a charcoal three-piece suit), holding a small soft bouquet of cream and blush florals between them. Background behind the couple is a simple two-tone soft sage and warm-cream gradient with subtle hand-drawn floral sprigs in the upper corners — minimal, modern, illustration-poster aesthetic, not photoreal. The lower 35% of the card is a clean cream banner with elegant centered typography: a small italic line "Together with their families" at the top, then large hand-lettered serif calligraphy "{NAMES}" as the visual centerpiece, then "request the honor of your presence" in romantic italic, then "{DATE}" in a clean small serif, then "{LOCATION}" in a smaller serif at the bottom. Soft watercolor paper texture across the whole card, visible imperfect brush edges, magazine-wedding-editorial polish. Portrait 5:7 composition.

**Rules the AI must follow:**
- Aspect ratio: 5:7 portrait — strict, locked at start and end
- Identity fidelity is the highest-priority visual constraint — preserve recognizable face shape, hair color and length, skin tone, and body proportions from the uploaded couple reference photo so the couple still looks like themselves, just stylized
- No Lorem Ipsum / no garbled text anywhere — the names, date, and location text must render exactly as specified, letter-for-letter, all spelled correctly
- Render the couple in soft-anime illustration style only — clean thin linework, smooth pastel shading, not photoreal, not 3D, not hyper-detailed pores
- Anatomically correct — exactly five fingers per hand, no extra limbs, natural hand position
- Single image output — one 5:7 portrait card, no contact sheet, no variant grid, no before/after
- Realistic imperfection in the paper texture and brush edges — no flat digital "AI plastic" finish on the card surface
- All text in English Latin script — no substitution to other writing systems
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit photo of the two of you together (front-facing, both faces visible) — without this the AI will invent a generic anime couple that doesn't look like you
- {NAMES} = Sarah Chen & Michael Park
- {DATE} = June 12, 2027
- {LOCATION} = The Manor House · Napa Valley · California

**Generate this image:**

Portrait 5:7 aspect ratio whimsical folk-art-illustrated wedding invitation card. The upper 70% of the card is a charming hand-drawn folk-art scene of the couple from the uploaded reference photo, in casual everyday outfits the couple would wear together (matching the uploaded photo's general style), inside the moment: {SHARED_MOMENT_SCENE}. The scene is filled with playful folk-art ornaments — vibrant floral patterns, small hearts and stars, embroidered borders, abstract decorative shapes, potted plants in the corners — rendered in a warm, hand-drawn aesthetic with visible pen and ink line texture. Color palette is rich and warm: dusty rose, mustard yellow, sage green, terracotta, cream paper background. The lower 30% of the card is a clean cream banner separated by a thin folk-art ornamental divider, with hand-lettered centered typography: a small line "Together we celebrate" at the top, then large folk-style hand-lettered "{NAMES}" as the visual centerpiece, then "are getting married" in romantic italic, then "{DATE}" in a clean small serif, then "{LOCATION}" in a smaller serif at the bottom. Subtle warm paper texture across the whole card, visible imperfect ink lines and slight asymmetry — every detail looks hand-drawn, never flat vector. Portrait 5:7 composition, charming storybook-folk-art wedding aesthetic.

**Rules the AI must follow:**
- Aspect ratio: 5:7 portrait — strict, locked at start and end
- Identity fidelity is the highest-priority visual constraint — preserve recognizable facial features, hair color and length, skin tone, and body proportions from the uploaded couple reference photo so the couple still looks like themselves stylized into folk-art
- No Lorem Ipsum / no garbled text anywhere — the names, date, and location text must render exactly as specified, letter-for-letter, all spelled correctly
- Render in hand-drawn folk-art illustration style only — visible pen and ink linework, slight asymmetry in the decorative motifs, warm paper texture; not photoreal, not 3D, not flat vector
- Anatomically correct — exactly five fingers per hand, no extra limbs, natural body proportions
- Single image output — one 5:7 portrait card, no contact sheet, no variant grid
- All text in English Latin script — no substitution to other writing systems
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear photo of the two of you together (both faces visible, casual everyday shot is fine) — without this the AI will invent a generic couple that doesn't look like you
- {SHARED_MOMENT_SCENE} = a small sunlit café table with two coffees and a shared pastry between them, a window with potted plants behind
- {NAMES} = Sarah Chen & Michael Park
- {DATE} = June 12, 2027
- {LOCATION} = The Manor House · Napa Valley · California

**Generate this image:**

Portrait 5:7 aspect ratio romantic sunset save-the-date wedding card. The composition is a real-photography sunset beach scene: two silhouettes of a couple standing close together at the water's edge of a sandy shore, looking out toward a calm ocean as the sun dips toward the horizon. The silhouettes occupy the lower 30-35% of the card (no facial detail visible — purely backlit shapes against the sky), with the sky filling the upper 65-70% as a stunning natural gradient of warm orange, soft coral pink, dusty rose, and a touch of deep mauve, the setting sun creating a brilliant gentle reflection across the calm waves, the faint outline of a distant coastline with rugged cliffs softly visible on the horizon. Upper-third typography overlay across the warm sky: "Save the Date" in elegant thin serif in cream-white, then large delicate hand-lettered serif "{NAMES}" as the visual centerpiece, then "{DATE}" in a clean small serif below, then "{LOCATION}" in a smaller serif at the very bottom of the typography zone, all centered, all in cream-white, leaving generous breathing room. Subtle warm film-grain texture, naturally imperfect color gradient, intimate and tranquil mood. Portrait 5:7 composition, real-photography sunset silhouette wedding aesthetic.

**Rules the AI must follow:**
- Aspect ratio: 5:7 portrait — strict, locked at start and end
- No Lorem Ipsum / no garbled text — the names, date, and location text must render exactly as specified, letter-for-letter, all spelled correctly, every character preserved
- The couple is rendered as silhouettes only — no facial features, no skin detail, no clothing detail, no identifiable hair texture, pure backlit dark shapes against the warm sky
- Real-photography sunset feel required — natural color gradient with warm-to-cool transitions, faint film grain, subtle natural haze; no AI-plastic flat gradient, no over-saturated digital sunset, no candy-bright color popping
- Sky negative space is the typography zone — keep it visually clean enough that the cream-white text reads clearly without competing detail
- Single image output — one 5:7 portrait card, no contact sheet, no variant grid
- All text in English Latin script — no substitution to other writing systems
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- {NAMES} = Sarah Chen & Michael Park
- {DATE} = June 12, 2027
- {LOCATION} = Tulum · Mexico

**Generate this image:**

A single photoreal 1:1 square LinkedIn studio headshot of the person in the uploaded reference image, shoulders-up framing, eyes meeting the camera with a soft natural closed-mouth smile, wearing {OUTFIT_STYLE}, against {BACKGROUND_STYLE}. Soft diffused studio key light from the upper-left with a gentle fill on the shadow side so skin tones stay clean and even, shallow depth of field, 85mm lens equivalent at f/2.8 for a flattering shoulders-up portrait, sharp focus on the eyes. The mood is polished, approachable, and modern-professional — confident but not stiff. Realistic skin texture preserved exactly as in the uploaded reference — visible pores, faint natural skin variation, no porcelain smoothing, no waxy AI-plastic finish. The person's face, bone structure, eye shape, nose, lips, hairline, and skin tone match the uploaded reference exactly so this is recognizably the same person. Output a single 1:1 square LinkedIn headshot, photoreal, identity-locked to the uploaded reference.

**Rules the AI must follow:**
- Aspect ratio 1:1 square — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's bone structure, eye shape, nose, lips, hairline, proportions, and skin tone exactly; the person must be unmistakably recognizable
- Realistic skin texture required — preserve visible pores, faint natural unevenness, light freckles or marks if present in the reference; no airbrushing, no porcelain smoothing, no glossy "AI plastic" beauty filter
- Eye-level camera angle, shoulders-up framing — no extreme low or high angles, no full-body, no head-only crop
- No glamour-magazine retouching, no heavy color grading — this is a professional profile photo, not an editorial cover; clean neutral color, true-to-life skin tone
- Single image output — no contact sheet, no variant grid, no before/after split
- No text, captions, logos, or watermarks anywhere in the image
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit front-facing photo of yourself (this is the face the AI will keep faithful — without it the headshot will not look like you)
- {OUTFIT_STYLE} = a dark navy blazer over a plain white shirt, no tie (or pick one — see Bonus tips below for swaps)
- {BACKGROUND_STYLE} = a clean soft-grey seamless studio backdrop with a subtle gradient (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single ultra-realistic 1:1 square executive corporate headshot of the person in the uploaded reference image, shoulders-up framing, calm confident expression with a subtle natural closed-mouth smile, wearing {OUTFIT_STYLE}, against {BACKDROP_STYLE}. Shot on an 85mm portrait lens at f/2 with a shallow depth of field — the eyes are tack-sharp and the backdrop falls into a soft gradient. Lighting is a soft diffused key light from the upper-left, gentle fill on the shadow side, and a subtle rim light separating the shoulders from the backdrop, giving a refined editorial-but-not-glamour quality. The person's face, bone structure, eye shape, nose, lips, hairline, and skin tone match the uploaded reference exactly so this is unmistakably the same person. Realistic skin texture preserved — visible pores, faint natural skin variation, no airbrushed porcelain finish. Cinematic color grading with clean neutral skin tones — slightly desaturated, slightly contrasty, premium and grounded, not Instagram-glossy. Output a single 1:1 square executive headshot, photoreal, identity-locked to the uploaded reference.

**Rules the AI must follow:**
- Aspect ratio 1:1 square — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's bone structure, eye shape, nose, lips, hairline, proportions, and skin tone exactly; the person must be unmistakably recognizable
- Camera spec is fixed — 85mm portrait lens equivalent at f/2, shallow depth of field, eye-level shoulders-up framing; do not substitute a wide angle or a flat-lay
- Realistic skin texture required — preserve visible pores, faint natural unevenness, light freckles or marks if present in the reference; no airbrushing, no porcelain smoothing, no glossy "AI plastic" beauty filter
- Cinematic but not editorial — subtle color grading and rim light are OK; no heavy fashion-magazine retouching, no aggressive contrast, no skin-blurring
- Single image output — no contact sheet, no variant grid, no before/after split
- No text, captions, logos, or watermarks anywhere in the image
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit front-facing photo of yourself (this is the face the AI will keep faithful — without it the headshot will not look like you)
- {OUTFIT_STYLE} = a tailored charcoal suit jacket over a crisp white dress shirt, slim dark tie (or pick one — see Bonus tips below for swaps)
- {BACKDROP_STYLE} = a soft gradient grey studio backdrop, slightly darker on one side for subtle depth (or pick one — see Bonus tips below for swaps)

**Generate this image:**

Generate one photo at a time from the set of four professional headshots below, all featuring the same person from the uploaded reference image. Each photo is a single 1:1 square image; do not combine them into a grid or contact sheet. Same person, different framing each time, same overall wardrobe family and same backdrop family so the four shots feel like one coordinated session.

- **Shot 1 — LinkedIn Primary**: A clean shoulders-up front-facing 1:1 square portrait of the person in the reference photo, eyes meeting camera, soft natural closed-mouth smile, wearing {OUTFIT_STYLE}, against {BACKDROP_STYLE}. Soft diffused key light from the upper-left, gentle fill. Photoreal, identity-locked, 1:1 square.
- **Shot 2 — Three-Quarter Shoulders-Up**: The same person turned slightly to the camera-left at a three-quarter angle, head turned slightly back toward camera, soft confident expression, same wardrobe family, same backdrop family, slightly more depth on one side from the directional light. Photoreal, identity-locked, 1:1 square.
- **Shot 3 — Side Profile**: The same person in a clean side profile portrait facing camera-left, calm composed expression looking forward into the frame (not at the camera), same wardrobe family, same backdrop family, soft natural light defining the jawline and cheekbone. Photoreal, identity-locked, 1:1 square.
- **Shot 4 — Wider Chest-Up Candid**: The same person in a slightly wider chest-up framing with a relaxed natural posture, slight off-center composition, warm authentic half-smile, eyes engaged with the camera, same wardrobe family, same backdrop family but with a softer ambient feel. Photoreal, identity-locked, 1:1 square.

**Rules the AI must follow (apply to every shot):**
- Aspect ratio 1:1 square — strict, applied to every single generation; stated at the start and the end of each shot description
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's bone structure, eye shape, nose, lips, hairline, proportions, and skin tone exactly; the same person must be unmistakably recognizable in all four shots
- Wardrobe and backdrop continuity — same outfit family and same backdrop family across all four shots so the set reads as one coordinated session, not four random photos
- Realistic skin texture required — preserve visible pores, faint natural unevenness, light freckles or marks if present in the reference; no airbrushing, no porcelain smoothing, no glossy "AI plastic" beauty filter
- Eye-level professional framing — no extreme low or high angles, no full-body, no editorial fashion poses
- Single image output per generation — one 1:1 photo per request; do not combine multiple photos into a grid, collage, or contact sheet
- No text, captions, logos, or watermarks anywhere in any of the photos
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit front-facing photo of yourself (this is the face the AI will keep consistent across all four shots — without it the photos will not look like you)
- {OUTFIT_STYLE} = a tailored navy blazer over a plain white shirt, no tie (or pick one — see Bonus tips below for swaps)
- {BACKDROP_STYLE} = a soft gradient grey studio backdrop with subtle depth (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single photoreal 1:1 square LinkedIn headshot of the person in the uploaded reference image, chest-up framing, eyes meeting the camera with a gentle natural soft smile, wearing {OUTFIT_STYLE}, photographed near a large window with soft diffused daylight pouring in from the camera-left, in {INDOOR_SETTING}. Light has a soft airy quality with gentle natural overexposure on the highlight side, low contrast, clean light-colored wall softly visible behind, shallow depth of field. The aesthetic is soft analog 35mm film — subtle film grain, slightly muted natural color, dreamy and understated but still grounded and professional, not artsy-editorial. The person's face, bone structure, eye shape, nose, lips, hairline, and skin tone match the uploaded reference exactly so this is unmistakably the same person. Realistic skin texture preserved — visible pores, faint natural skin variation, soft natural makeup look at most, no airbrushed porcelain finish, no AI-plastic beauty filter. Output a single 1:1 square natural-light headshot, photoreal, identity-locked to the uploaded reference.

**Rules the AI must follow:**
- Aspect ratio 1:1 square — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's bone structure, eye shape, nose, lips, hairline, proportions, and skin tone exactly; the person must be unmistakably recognizable
- Natural window-light look — soft directional daylight from camera-left, gentle highlight roll-off, no harsh studio strobes, no ring light flatness
- Realistic skin texture required — preserve visible pores, faint natural unevenness, light freckles or marks if present in the reference; no airbrushing, no porcelain smoothing, no glossy "AI plastic" beauty filter
- Soft analog film aesthetic — subtle film grain, gentle natural color, low contrast; not heavy fashion-magazine retouching, not Instagram-glossy
- Eye-level chest-up framing — no full-body, no extreme low or high angles, no editorial fashion poses
- Single image output — no contact sheet, no variant grid, no before/after split
- No text, captions, logos, or watermarks anywhere in the image
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit front-facing photo of yourself (this is the face the AI will keep faithful — without it the headshot will not look like you)
- {OUTFIT_STYLE} = a soft oversized white button-up shirt, smart-casual and unfussy (or pick one — see Bonus tips below for swaps)
- {INDOOR_SETTING} = a minimal indoor space near a large window with white sheer curtains, clean light-colored wall behind (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single photoreal 3:4 portrait US-style professional ID photo of the exact person from the uploaded reference image — match their face, bone structure, eyes, nose, lips, hair, and skin tone exactly, 100% face match, no facial changes. The person is centered horizontally, head and shoulders only, with the head from top of head to chin occupying 70-80% of the vertical frame height. Background is pure white #FFFFFF, completely uniform, no gradient, no shadow on the face, no shadow on the background. Lighting is even diffuse studio-style — soft frontal light, no harsh shadows, no rim lighting, moderate contrast for biometric legibility, realistic natural skin texture with visible pores and fine micro-detail (no AI-plastic smoothing). The person wears {OUTFIT}, sits squarely facing the camera, neutral expression, mouth closed, no smile, both eyes open and looking straight at the camera. Camera: 85mm equivalent lens, f/8, ISO 100, eye-level angle, no perspective distortion. Single 3:4 portrait ID photo, head-and-shoulders, pure white background, federal-ID-grade clarity.

**Rules the AI must follow:**
- Aspect ratio 3:4 portrait — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's face, bone structure, eyes, nose, lips, hair, and skin tone exactly; 100% face match, no facial changes, no beautification, no age adjustment
- Background must be pure white #FFFFFF, completely uniform — no gradient, no texture, no shadow on face, no shadow on background
- Neutral expression required — mouth closed, no smile, no teeth visible, both eyes open and looking straight at the camera
- No glasses, no hat, no head covering, no hair covering the face, no hair covering the eyebrows, shoulders included at the bottom of the frame and not cropped
- Realistic natural skin texture required — visible pores, faint freckles or skin micro-detail, slight micro-asymmetry; no over-retouched look, no porcelain skin, no AI-plastic smoothing
- Single image output — no contact sheet, no variant grid, no before/after split, no multiple angles in one frame
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit front-facing photo of yourself (this is the face the AI will keep — without it the photo will not be you)
- {OUTFIT} = plain dark blazer or shirt collar visible at the bottom of the frame, no patterns, no logos (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single photoreal 3:4 portrait scanned-print ID photo of the exact person from the uploaded reference image — match their face, bone structure, eyes, nose, lips, hair, and skin tone exactly, 100% face match, no facial changes. The person is centered, head and shoulders only, head from top of head to chin occupying 70-80% of the vertical frame height. Background is a smooth uniform {BACKDROP_COLOR} photographic backdrop with very subtle paper texture. Lighting is direct on-camera flash — even bright exposure across the face, small bright catchlight reflections in both eyes, mild semi-gloss print shine with slight glare, fine paper grain visible across the whole image, subtle ink warmth in the shadows. The person wears {OUTFIT}, sits squarely facing the camera, neutral closed-mouth expression or a very subtle pleasant micro-smile, mouth closed, both eyes open and looking straight at the camera. Realistic natural skin texture — visible pores, faint freckles or skin micro-detail, no plastic smoothing. The overall look is "this was photographed at the front office and printed on a school ID card" — scanned-print authenticity, not a clean studio digital portrait. Single 3:4 portrait scanned-print ID photo, head-and-shoulders, colored backdrop, paper-grain finish.

**Rules the AI must follow:**
- Aspect ratio 3:4 portrait — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's face, bone structure, eyes, nose, lips, hair, and skin tone exactly; 100% face match, no facial changes
- Scanned-print aesthetic is required — direct on-camera flash exposure, semi-gloss print shine, fine paper grain across the whole image, subtle ink warmth in shadows; do not render as a clean digital studio portrait
- Neutral or subtle micro-smile only — mouth closed, no teeth showing, both eyes open and looking straight at the camera
- No glasses, no hat, no head covering, no hair covering the face, shoulders included at the bottom of the frame and not cropped
- Realistic natural skin texture required — visible pores, faint freckles, slight micro-asymmetry; no over-retouched look, no porcelain skin, no AI-plastic smoothing
- Single image output — no contact sheet, no variant grid, no before/after split, no multiple angles in one frame
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit front-facing photo of yourself (this is the face the AI will keep — without it the photo will not be you)
- {BACKDROP_COLOR} = cool school-photo blue (or pick one — see Bonus tips below for swaps)
- {OUTFIT} = a navy blazer over a crisp white collared shirt (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single ultra-realistic {RATIO} passport-size photo of me with exact face likeness from the uploaded reference image — 100% face match, no facial changes, no beautification, no age adjustment, matching the bone structure, eyes, nose, lips, hair, and skin tone exactly. This photo is for {COUNTRY} passport / visa / national ID submission and must follow that country's biometric spec. The person is centered horizontally, head and shoulders only, head from top of head to chin occupying 70-80% of the vertical frame height, shoulders included at the bottom of the frame and not cropped. Background is pure white #FFFFFF, completely uniform, no gradient, no shadow on the face, no shadow on the background. Lighting is even diffuse studio-style — soft frontal light, no harsh shadow, no rim lighting, moderate contrast for biometric legibility. The person wears a plain dark top with collar visible at the bottom of the frame, no patterns, no logos, no jewelry visible. Neutral expression, mouth closed, no smile, no teeth visible, both eyes open and looking straight at the camera. Realistic natural skin texture — visible pores, faint freckles or skin micro-detail, slight micro-asymmetry, no plastic smoothing. Camera: 85mm equivalent lens, f/8, ISO 100, eye-level angle, no perspective distortion. Single {RATIO} passport photo, head-and-shoulders, pure white background, biometric-spec compliant.

**Rules the AI must follow:**
- Aspect ratio {RATIO} — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's face, bone structure, eyes, nose, lips, hair, and skin tone exactly; 100% face match, no facial changes, no beautification, no age adjustment, no slimming
- Background must be pure white #FFFFFF, completely uniform — no gradient, no texture, no shadow on face, no shadow on background
- Neutral expression required — mouth closed, no smile, no teeth visible, both eyes open and looking straight at the camera
- No glasses, no hat, no head covering, no hair covering the face, no hair covering the eyebrows, shoulders included at the bottom of the frame and not cropped
- Realistic natural skin texture required — visible pores, faint freckles or skin micro-detail, slight micro-asymmetry; no over-retouched look, no porcelain skin, no AI-plastic smoothing
- Single image output — no contact sheet, no variant grid, no before/after split, no multiple angles in one frame
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit front-facing photo of yourself (this is the face the AI will keep — without it the photo will not be you)
- {COUNTRY} = United States (or pick one — see Bonus tips below for ratios per country)
- {RATIO} = 1:1 square (US 2×2 inch — or pick the ratio matching your country, see Bonus tips below)

**Generate this image:**

A single photoreal 3:4 portrait corporate ID photo of the person in the uploaded reference image, head and shoulders only, framed for a workplace badge / HR system / tenant directory / gig-platform verification upload. The person is centered horizontally, head from top of head to chin occupying 65-75% of the vertical frame, shoulders visible at the bottom of the frame and not cropped. Eyes meeting the camera with an {EXPRESSION_TONE}, modern-professional posture. They wear {OUTFIT}, against a {BACKDROP_TONE} seamless backdrop — softly graduated, no pure white, no harsh studio glare. Soft front lighting with a gentle natural falloff to the shadow side — warm, even, slightly directional from camera-left; gentle natural shadow on the cheek-side is OK and adds dimension. The mood is approachable-professional and trustworthy — the photo HR or a background-check reviewer wants to see, not a mugshot, not a glamour shot. Realistic skin texture preserved exactly as in the uploaded reference — visible pores, faint natural unevenness, no porcelain smoothing, no AI-plastic finish. The person's face, bone structure, eye shape, nose, lips, hairline, and skin tone match the uploaded reference exactly so this is unmistakably them. Single 3:4 portrait corporate ID photo, head-and-shoulders, soft corporate-neutral backdrop, warm-professional tone.

**Rules the AI must follow:**
- Aspect ratio 3:4 portrait — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's bone structure, eye shape, nose, lips, hairline, proportions, and skin tone exactly; no beautification, no slimming, no age adjustment
- Backdrop is a soft corporate-neutral tone — softly graduated, not pure white (#FFFFFF) and not federal-ID flat-white; use {BACKDROP_TONE} as specified
- Approachable-professional expression — soft natural closed-mouth smile or calm confident neutral expression as specified in {EXPRESSION_TONE}; this is NOT a biometric passport photo, smile is permitted
- Realistic skin texture required — visible pores, faint natural unevenness, slight micro-asymmetry; no porcelain smoothing, no waxy AI-plastic surface, no beautify filter
- Soft front lighting with gentle natural falloff — slight directional shadow on the cheek-side is OK and welcome; no harsh biometric flat-flash, no editorial dramatic side-light
- Single image output — no badge / lanyard / employee-number overlay drawn onto the photo, no contact sheet, no variant grid, no before/after split
- No text, captions, logos, employee numbers, or watermarks rendered into the image — the destination HR / badge system will overlay those at template-render time
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit front-facing photo of yourself (this is the face the AI will keep — without it the badge photo will not look like you)
- {EXPRESSION_TONE} = soft natural closed-mouth smile, warm and approachable, eyes meeting the camera (or pick one — see Bonus tips below)
- {OUTFIT} = a clean business-casual top — a plain solid-color collared shirt or simple knit in a neutral tone, top button closed, no patterns, no logos (or pick one — see Bonus tips below)
- {BACKDROP_TONE} = soft office-blue (#D7E0EA), softly graduated, corporate-neutral (or pick one — see Bonus tips below)

**Generate this image:**

A single photoreal 4:5 portrait editorial-fashion collage canvas of the same person from the uploaded reference image, shown in 8 distinct full-body looks, arranged organically across one clean {BACKDROP_COLOR} studio backdrop — same face, same bone structure, same proportions across all 8 figures. Each of the 8 figures stands or sits in a relaxed natural posture, drawn from {OUTFIT_THEME}, with visible variation in outfit silhouette, color, and styling so the 8 looks read as a coordinated multi-context wardrobe rather than 8 versions of the same outfit. Figures are scaled and overlapped naturally — some full-height, some chest-up — placed at varied heights across the canvas with soft natural shadows underneath, not stacked in rows. Subtle handwritten-style arrows and small lowercase pencil annotations point to key wardrobe pieces (e.g. "blazer," "loafers," "denim"). Soft diffused studio daylight, realistic skin texture with visible pores and natural micro-imperfection, editorial-magazine aesthetic, professionally art-directed feel. Single 4:5 portrait editorial-collage canvas with the same person in 8 outfits.

**Rules the AI must follow:**
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's bone structure, eyes, nose, lips, proportions, and skin tone exactly; the same person must be recognizable in all 8 figures with no drift across the canvas
- The 8 figures are arranged organically across one freeform editorial canvas — no grid, no panels, no borders, no boxed layout, no contact-sheet feel (this is what separates the editorial collage from a phone collage app)
- Each of the 8 outfits must be visibly distinct in silhouette, color palette, and styling — no two looks should read as variations of each other
- Realistic skin texture required — visible pores, faint natural unevenness, slight micro-asymmetry; no porcelain skin, no AI plastic smoothing, no over-retouched magazine beauty filter
- Handwritten arrows and annotations are subtle accents — thin pencil-style lines, small lowercase labels, sparse placement; do not dominate the canvas, do not turn it into a Pinterest moodboard
- Single image output — one editorial-collage canvas, not a 8-panel grid or contact sheet of separate photos
- No text other than the small handwritten wardrobe-piece labels described above — no headlines, no captions, no watermarks, no logos
- All text in English Latin script if any incidental signage appears
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit front-facing photo of yourself (this is the face the AI will keep consistent across all 8 looks — without it the figures will not look like you)
- {OUTFIT_THEME} = a mixed personal-brand wardrobe spanning smart-casual office, podcast-host casual, event-speaking blazer, weekend creator, fitness, polished business-casual, casual all-black creator, and one statement piece (or pick one — see Bonus tips below for swaps)
- {BACKDROP_COLOR} = warm cream studio paper (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single photoreal 9:16 vertical cinematic three-panel collage featuring the same person from the uploaded reference image performing {POSE_SEQUENCE} in a calm, minimal studio environment, wearing {OUTFIT}. The three panels are stacked vertically inside one continuous canvas — same person, same face, same bone structure, same proportions in all three panels with absolutely no drift. Soft diffused natural daylight from one direction creates a serene atmosphere with gentle realistic shadows. Minimal uncluttered studio background, shallow depth of field on the subject, realistic skin texture with visible pores and faint natural imperfection, editorial-wellness photography aesthetic, cinematic color grading, hyper-realistic, slightly zoomed-in framing on each pose. Single 9:16 vertical three-panel cinematic collage of the same person.

**Rules the AI must follow:**
- Aspect ratio 9:16 vertical — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's face 100% across all three panels (same bone structure, same eyes, same nose, same lips, same proportions, same skin tone); the same person must be unmistakably recognizable in every panel
- The three panels are clearly distinct poses — no duplicated poses, no near-identical body positions, each panel reads as a different moment of the same practice
- Realistic skin texture required — visible pores, faint natural unevenness, natural muscle definition; no porcelain skin, no AI plastic smoothing, no over-retouched fitness-magazine beauty filter
- Single image output — one continuous three-panel vertical canvas, not three separate images and not a four-square grid
- No text, captions, watermarks, or logos anywhere in the image
- All text in English Latin script if any incidental signage appears in the background
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit front-facing photo of yourself (this is the face the AI will keep consistent across all three panels — without it the poses will not look like you)
- {POSE_SEQUENCE} = three yoga poses — top panel: tree pose, standing balanced with hands pressed together above the head, calm focused expression. middle panel: downward dog, body forming a clean inverted V shape, soft natural light highlighting the form. bottom panel: cross-legged meditation pose, eyes gently closed, hands resting on knees, peaceful relaxed expression (or pick one — see Bonus tips below for swaps)
- {OUTFIT} = a deep burgundy fitted activewear set (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single 1:1 square 3x3 grid of 9 stylized 3D character avatars rendered in {ANIMATION_STYLE}, all featuring the same character based on the person from the uploaded reference image — recognizably them in stylized form (same face geometry, same hair color and shape, same overall facial proportions). The 9 panels show {EXPRESSION_SET}, with each panel framing the character chest-up against a different soft solid pastel background color. The character wears a simple clean neutral-color top across all 9 panels — no logos, no branded merch, no distracting clothing details. Soft studio lighting per panel, expressive readable eyes and clear facial expressions, smooth high-quality 3D rendering with subtle imperfection in shading (not flat plastic), modern emoji/reaction-pack character design, clean composition, ultra-detailed, vibrant but not oversaturated colors. 9 equal-sized square panels arranged in a clean 3x3 grid with thin neutral gutters between panels — each panel must be independently readable when cropped to a small circular avatar against a dark-mode UI. Single 1:1 square 3x3 emoji-reaction grid of the same stylized character.

**Rules the AI must follow:**
- Aspect ratio 1:1 square — strict, locked at the start and the end of the prompt
- Character identity preservation is the highest-priority constraint — the stylized character must be recognizably the same person from the uploaded reference photo in all 9 panels (same face geometry, same hair, same overall proportions); no drift between panels, no different-person variations
- All 9 panels render in the same animation style and same rendering quality — no mixing of Pixar with 2D vector with claymation in the same grid
- Each panel must work as a standalone tiny avatar — silhouette must be clear, background must be a SOLID pastel (not gradient, not patterned), character must remain readable when the panel is cropped to a small circular icon against a dark UI
- Subtle realistic imperfection in the rendering — natural light gradation on the face, slight asymmetry in expression, no flat plastic "AI animation" look
- Single image output — one 3x3 grid canvas, not 9 separate images, not a contact sheet of style options
- No text, captions, watermarks, or logos anywhere in the grid
- All text in English Latin script if any incidental signage appears
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit front-facing photo of yourself (this is the face the AI will stylize into the same character across all 9 panels — without it the avatars will not look like you)
- {ANIMATION_STYLE} = modern Pixar-style 3D animation with warm, soft lighting and expressive eyes (or pick one — see Bonus tips below for swaps)
- {EXPRESSION_SET} = a newsletter/community reaction set — panel 1: bright excited smile / panel 2: surprised lightbulb-moment eyes wide / panel 3: gentle thinking, hand to chin / panel 4: confident thumbs-up with a half-smile / panel 5: warm laughing with eyes crinkled / panel 6: facepalm self-deprecating grin / panel 7: skeptical raised eyebrow / panel 8: celebrating hands-in-air / panel 9: camera-ready calm friendly headshot (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single 4:5 portrait scrapbook-aesthetic composition built on top of the uploaded reference image. Strictly preserve the original photo from the upload — the main subject's face, pose, outfit, lighting, background, and all original details must remain 100% unchanged and unedited. Do not recreate, do not regenerate, do not re-render the original photo; only ADD new elements on top of it. Onto this preserved photo, add 3-5 small "mini version" chibi-style 3D collectible-figurine characters of the same person, placed naturally around the scene (perched on objects, sitting on the desk, on a shoulder, beside a coffee mug). The mini chibi figures must share the EXACT same face geometry, hairstyle, outfit, and overall look as the main subject in the original photo — they are tiny alter-egos, not different characters. The mini figures are shown {MINI_AVATAR_ACTIVITIES}. Then overlay subtle handwritten pencil-style line annotations across parts of the image: thin arrows, small stars, sparkles, and short lowercase labels in a {ANNOTATION_PALETTE} palette — sparse placement, never covering more than 30% of the canvas surface. Warm cozy lighting matched to the original photo, dreamy scrapbook aesthetic, soft depth of field on the minis, polished but playful. Single 4:5 portrait scrapbook canvas — original photo preserved + chibi mini alter-egos + handwritten annotations on top.

**Rules the AI must follow:**
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt
- Original photo preservation is the highest-priority constraint — the uploaded photo's main subject (face, pose, expression, outfit), the lighting, and the background must remain 100% unchanged; do NOT recreate, regenerate, or re-render the original photo, only ADD new elements on top of it
- Chibi mini-figure identity lock — the mini alter-egos must share the EXACT face geometry, hairstyle, and outfit of the main subject in the original photo (they are tiny versions of the same person, not different characters)
- 3 to 5 mini figures only — natural varied placement, not clustered, not symmetrical, scaled small enough to feel like alter-egos perched in the scene
- Handwritten annotations stay subtle — pencil-line aesthetic, never overpower the original photo, never cover more than ~30% of the canvas
- Realistic skin texture preserved on the main subject — keep all original pores, freckles, and micro-imperfection from the uploaded photo verbatim; only the chibi minis are stylized, the human subject remains photoreal
- Single image output — one scrapbook canvas, not a before/after split, not multiple variants
- No printed text other than the small handwritten lowercase labels described above — no headlines, no watermarks, no logos, no captions
- All text in English Latin script
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a real photo of you in your work environment — at your desk, behind a podcast mic, on stage, in your studio (this photo will be preserved unchanged; the AI only adds chibi minis and annotations on top — without it the prompt cannot work)
- {MINI_AVATAR_ACTIVITIES} = doing different small creator activities — one writing in a tiny journal, one sitting beside a coffee mug, one talking into a mini podcast mic, one waving to the camera (or pick one — see Bonus tips below for swaps)
- {ANNOTATION_PALETTE} = soft warm neutrals — cream base with sage green, dusty terracotta, and warm sand accents (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single photoreal 2:3 vertical cinematic editorial portrait of the person from the uploaded reference image — match their bone structure, eyes, nose, lips, proportions, and skin tone exactly so the subject is clearly recognizable. The subject is seated upright with quiet authority, body angled three-quarters to camera, both hands resting naturally on the chair arms, eyes meeting camera with a calm, grounded, slightly confident expression — not smiling, not stern. They wear {ATTIRE_STYLE}. The setting is {SETTING}, with a soft directional cinematic key light from the upper-left at roughly 45° and a gentle cool fill on the shadow side, producing rich but never harsh contrast. Three-quarter-length composition — the subject is visible from the top of the head down to mid-thigh, with both hands fully in frame resting on the chair arms; the head occupies approximately 1/7 of frame height (editorial portrait standard), and head, hands, and torso are rendered at correct anatomical scale relative to one another, anchored slightly above center. The background remains atmospheric and out of focus but never competes with the subject. Skin shows visible pores, fine micro-texture, and natural micro-asymmetry — no porcelain smoothing. Shot on an 85mm lens at f/1.4, shallow depth of field, premium editorial photography style; 85mm telephoto compression keeps the head and body in true anatomical proportion — no enlarged head, no shrunken or misshapen hands, no wide-angle facial distortion. Single 2:3 vertical cinematic editorial portrait, identity-locked to the uploaded reference.

**Rules the AI must follow:**
- Aspect ratio 2:3 vertical portrait — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's bone structure, eyes, nose, lips, proportions, and skin tone exactly; the subject must be unmistakably the same person
- Three-quarter-length crop — frame includes from top of head down to mid-thigh, with both hands fully visible on the chair arms; head occupies approximately 1/7 of frame height; head, hands, and torso must remain in true anatomical proportion — no enlarged head, no shrunken or misshapen hands, no wide-angle facial distortion
- Realistic skin texture required — visible pores, fine micro-asymmetry, natural unevenness; no porcelain smoothing, no over-retouched beauty filter, no waxy AI-plastic surface
- One human figure only — solo subject; no advisors, no crowd, no background people, no multi-exposure ghosts
- No text, captions, watermarks, logos, brand marks, or readable signage anywhere in the frame
- Single image output — no contact sheet, no variant grid, no before/after split, no multiple angles in one frame
- All text in English Latin script if any incidental signage appears
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit front-facing photo of your face (this is the face the AI will keep — without it the portrait will not look like you)
- {SETTING} = a modern minimalist studio with a softly lit warm-grey paper backdrop and a single architectural chair (or pick one — see Bonus tips below for swaps)
- {ATTIRE_STYLE} = a well-tailored charcoal suit jacket over a crisp white open-collar shirt, no tie, modern minimal silhouette (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single ultra-realistic, cinematic 2:3 vertical launch-hero portrait of the person from the uploaded reference image — match their bone structure, eyes, nose, lips, proportions, and skin tone exactly; the face must stay fully intact and clearly recognizable. The subject is shown three-quarter body, facing slightly off-camera in a forward-arriving pose, with a calm, resolute, slightly forward-looking expression and eyes confidently toward the horizon. They wear a well-tailored modern outfit with a sharply cut silhouette and a quietly powerful presence. The lower torso and one trailing arm begin to dissolve into luminous particles as {METAPHOR_EFFECT}, but the face, head, shoulders, and upper body remain fully solid and fully recognizable. Behind and partially surrounding the subject is a massive glowing backlight ring acting as a strong rim and silhouette light. The setting is minimal and cinematic — a desolate concrete surface beneath an atmospheric overcast sky. The {COLOR_PALETTE}. Shallow depth of field, glow/bloom rim lighting, subtle particle dispersion, photo-real digital painting aesthetic, dramatic cinematic composition, ultra-detailed skin and fabric textures. Single 2:3 vertical launch-hero portrait, identity-locked, face fully intact.

**Rules the AI must follow:**
- Aspect ratio 2:3 vertical portrait — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's bone structure, eyes, nose, lips, proportions, and skin tone exactly; the face must remain fully solid and clearly recognizable; never let the dissolving effect touch the face, head, or shoulders
- Realistic skin texture required — visible pores, fine micro-asymmetry, natural unevenness; no porcelain smoothing, no over-retouched beauty filter, no waxy AI-plastic surface
- No readable text anywhere in the frame — no book titles, no course names, no dates, no logos, no captions, no watermarks (these will hallucinate as gibberish — overlay any text separately in Canva or Figma)
- Single image output — no contact sheet, no variant grid, no before/after split, no multiple angles in one frame
- All text in English Latin script if any incidental signage appears
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit front-facing photo of your face (this is the face the AI will keep — without it the portrait will not look like you)
- {METAPHOR_EFFECT} = thick cloud-like grey smoke filled with glowing golden dust particles, suggesting active transformation and emergence (or pick one — see Bonus tips below for swaps)
- {COLOR_PALETTE} = color palette is dominated by monochrome greys, deep slate blues, and soft whites, contrasted with intense amber-gold glow from the backlight (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single ultra-realistic high-fashion 3:4 editorial portrait of the person from the uploaded reference image — match their bone structure, eyes, nose, lips, proportions, and skin tone exactly so the subject is clearly recognizable. The subject stands confidently in a monochromatic {PALETTE_COLOR} urban environment, wearing a well-tailored modern outfit in the same {PALETTE_COLOR} palette, subtle natural makeup and soft matte skin texture, hair styled naturally and slightly catching the breeze. Beside and slightly behind them is {SIGNATURE_OBJECT}, partially wrapped in soft swirling atmospheric fog — the only meaningful object in the frame other than the subject. The environment is almost entirely the same monochrome {PALETTE_COLOR} — same-tone architecture, same-tone fog, same-tone lighting — creating a clean editorial composition. The subject occupies 55-70% of the vertical frame, anchored slightly off-center; the signature object reads as a clear secondary focal point. Shot on an 85mm lens, shallow depth of field, soft cinematic directional lighting, studio-quality color grading, real-world textures, Vogue magazine aesthetic, modern editorial photography. Single 3:4 vertical editorial portrait, identity-locked, photoreal.

**Rules the AI must follow:**
- Aspect ratio 3:4 vertical portrait — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's bone structure, eyes, nose, lips, proportions, and skin tone exactly; the subject must be unmistakably the same person
- The signature object is rendered with anatomical and physical realism — proportions, weight, surface details, and edges must match how the real object exists in the world; no extra heads, no impossible curves, no impossible repetitions
- Realistic skin texture and fabric texture required — visible pores, fine micro-asymmetry, authentic fabric folds, natural hair strands; no porcelain smoothing, no over-retouched filter, no waxy AI-plastic surface
- No fantasy elements, no CGI appearance, no cartoon style, no anime aesthetic — photoreal editorial photography only
- No readable text anywhere — no captions, no logos, no watermarks, no readable signage in the background
- Single image output — no contact sheet, no variant grid, no before/after split
- All text in English Latin script if any incidental signage appears
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit front-facing photo of your face (this is the face the AI will keep — without it the portrait will not look like you)
- {PALETTE_COLOR} = cream — soft warm cream tones throughout the scene, with subtle ivory and pale-beige variations (or pick one — see Bonus tips below for swaps)
- {SIGNATURE_OBJECT} = a stack of three well-worn hardcover books, spines visible, held casually at the subject's side (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single cinematic photoreal 1:1 square aesthetic portrait of the person from the uploaded reference image — match their bone structure, eyes, nose, lips, proportions, and skin tone exactly so the subject is unmistakably recognizable. The subject is shown chest-up, standing near a window with soft warm side light gently catching one side of their face, in a natural candid pose — slight relaxed smile or quiet thoughtful expression, eyes either meeting the camera or looking slightly off to the side. They wear {CLOTHING_STYLE}, with natural fabric texture and authentic folds. The composition is clean and Instagram-inspired with shallow depth of field, slightly off-center subject placement, ultra-realistic skin texture, soft window-side warm light, gentle film grain, and a calm dreamy atmosphere. Around the subject, layered on top of the photo, are soft hand-drawn pastel doodles and artistic scribble elements — {DOODLE_THEME} — drawn with the looseness of a real Procreate sketch, slightly imperfect lines, semi-transparent pastel colors, scattered organically so they decorate but do not block the subject's face. Single 1:1 square aesthetic portrait, identity-locked, photo with decorative doodle overlay.

**Rules the AI must follow:**
- Aspect ratio 1:1 square — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's bone structure, eyes, nose, lips, proportions, and skin tone exactly; the subject must be unmistakably the same person; doodles must never overlap or obscure the face
- Realistic skin texture required — visible pores, fine micro-asymmetry, natural unevenness; no porcelain smoothing, no over-retouched filter, no waxy AI-plastic surface
- Doodles must remain decorative shapes and icons only — no readable words, no readable phrases, no chat-bubble text, no handwritten lettering (AI will render those as gibberish — overlay any actual text separately afterward in Canva or Figma)
- Doodles look hand-drawn — slightly imperfect lines, varied stroke weight, semi-transparent soft pastel colors; never crisp vector graphics, never clipart
- Single image output — no contact sheet, no variant grid, no before/after split
- All text in English Latin script if any incidental signage appears
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit front-facing photo of your face (this is the face the AI will keep — without it the portrait will not look like you)
- {CLOTHING_STYLE} = an oversized neutral-toned streetwear top — cream or soft beige heavyweight cotton, simple, modern, slightly relaxed fit (or pick one — see Bonus tips below for swaps)
- {DOODLE_THEME} = playful arrows, stars, smiley faces, swirls, sketch outlines, tiny hearts, and small abstract scribbles for a friendly creative vibe (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single 3:4 portrait image laid out as a vertical 3x1 webcam photobooth strip, all three panels of the exact same person from the uploaded reference image — preserve their face, eye shape and color, nose, lip shape, hair color and texture, and skin tone across all three panels so they are unmistakably the same person. Each panel is a low-resolution warm-toned webcam still with a slightly grainy / noisy Y2K aesthetic, framed tight from the chest up, fills its panel edge-to-edge with no inner border, shot at the slightly tilted handheld angle of an early-2000s webcam. The screen light is warm and dim, with a soft cozy bedroom background that stays softly out of focus. The person wears a casual {OUTFIT_STYLE} that fits the {VIBE} mood. Panel 1: a soft natural closed-mouth smile, eyes meeting the lens, hand relaxed by the cheek. Panel 2: a playful smirk with one eyebrow slightly raised, head tilted, eyes glancing slightly off-camera. Panel 3: a candid laugh caught mid-moment, eyes crinkled, very natural. Three panels stacked vertically, thin tight gutters between them, no captions or text anywhere, no outer border. Single 3:4 portrait image showing the 3x1 Y2K webcam strip, identity-locked across all three panels.

**Rules the AI must follow:**
- Aspect ratio 3:4 portrait overall — strict, locked at the start and the end of the prompt; the strip is 3 stacked panels inside this single image
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's face shape, eye shape and color, nose, lip shape, hair, and skin tone exactly; the same person must be recognizable in all three panels
- Y2K webcam realism required — slight grain, slight low-resolution softness, warm dim screen-light color cast; no DSLR studio sharpness, no modern smartphone HDR, no professional beauty filter
- Realistic skin texture required — visible pores, faint freckles, slight micro-asymmetry; no over-retouched look, no porcelain skin, no AI plastic smoothing
- No text, no captions, no watermarks, no panel labels, no timestamps anywhere in the image
- Single image output — exactly one 3:4 image containing the vertical 3x1 strip; no separate panels, no contact sheet, no extra grid options
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears in the background

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear front-facing selfie of yourself in good light (this is the face the AI will keep consistent across all three panels — without it the strip will not look like you)
- {OUTFIT_STYLE} = a fitted black off-shoulder top (or pick one — see Bonus tips below for swaps)
- {VIBE} = Y2K bedroom-cozy with a slightly nostalgic early-2000s feel (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single 1:1 square photoreal image that keeps the uploaded reference photo of the person unchanged as the base — same framing, same lighting, same composition, same face exactly. Around them, add several small "chibi 3D" miniature versions of the SAME person (preserve their face, hair color and texture, and skin tone so every chibi reads as clearly the same person, just tiny and stylized), each chibi about the size of a coffee cup, scattered around the frame in believable interactions with the scene: one sitting on the person's shoulder, one climbing up their sleeve, one perched on a nearby surface holding a tiny coffee, one lounging against an object reading a tiny book. Each chibi has soft Pixar-style 3D rendering with big expressive eyes and a friendly cozy expression. Realistic shadows and depth so the chibis feel physically present in the same lighting as the base photo, not pasted on. Add one small handwritten note floating gently near the person, exactly the text {CAPTION} written in soft script — keep it short, English Latin script only, no Lorem Ipsum, no gibberish. Overall mood is cinematic, cozy, slightly dreamy, viral-IG aesthetic. Single 1:1 square photoreal image with chibi mini-me effect, identity-locked base + identity-locked chibis.

**Rules the AI must follow:**
- Aspect ratio 1:1 square — strict, locked at the start and the end of the prompt (works for IG profile crop + Pinterest re-crop)
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's face, eye shape and color, nose, lip shape, hair, and skin tone exactly for BOTH the base person AND every chibi mini-version; do not invent a different person, do not drift in hair color or face shape across the chibis
- Base photo of the person stays unchanged — same framing, same lighting, same pose, same face; only ADD chibis around them, do not restyle the base photo itself
- Realistic skin and shadow texture required — visible pores and natural micro-detail on the base person; chibis cast believable shadows that match the base photo's light direction; no AI-plastic surface on either
- Render the {CAPTION} text exactly as specified, all spelled correctly, in English Latin script — no Lorem Ipsum, no garbled letters, no extra captions, no watermarks, no logos
- Single image output — one 1:1 photoreal image, no before/after split, no character sheet, no separate chibi grid alongside
- Output the image directly without explaining the prompt back
- All text in English Latin script

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear photo of yourself (preferably half-body or full-body with some space around you for the chibis to live in) — this is the face the AI will preserve on you AND on every tiny chibi version of you
- {CAPTION} = Little versions of me, living my quiet moments (or pick a short line — see Bonus tips below for swaps)

**Generate this image:**

A single 1:1 square ultra-realistic smartphone mirror selfie of the exact person from the uploaded reference image — match their face, eye shape and color, nose, lip shape, hair color and texture, body proportions, and skin tone exactly so they are recognizably the same person, not a generic AI model and not a stylized version. The person is standing in front of a {LOCATION} mirror, holding a modern smartphone naturally to capture the shot — phone partially visible covering a small section of the upper face is fine, in fact preferred for realism — wearing a casual {OUTFIT_STYLE}, with a soft natural relaxed expression (slight closed-mouth smile or neutral, not a posed beauty-shot smile). Phone-camera realism: visible skin texture (real pores, faint freckles, slight micro-asymmetry, subtle natural shine), slightly imperfect handheld framing, soft ambient light from the room (no studio ring light), believable mirror reflection physics including the mirror frame edge and natural background depth visible behind the person. Background is the actual {LOCATION} environment, softly out of focus but coherent — clearly the same physical space, not a studio cyclorama. Feels like a real casual IG post by an actual person, indistinguishable from a real smartphone photo. Single 1:1 square ultra-realistic mirror selfie, identity-locked to the uploaded person, IG-profile crop ready.

**Rules the AI must follow:**
- Aspect ratio 1:1 square — strict, locked at the start and the end of the prompt (IG profile crops to a circle inside this square)
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's face shape, eye shape and color, nose, lips, hair, body proportions, and skin tone EXACTLY; this is the same person posting a selfie, not a generic model with their face
- Realistic skin texture required — visible pores, faint freckles, slight micro-asymmetry, subtle natural shine; absolutely no airbrushing, no porcelain skin, no AI-plastic smoothing, no over-retouched glossy look
- Phone-camera handheld realism — slight natural framing imperfection is OK and preferred, soft ambient room light only; no DSLR studio lighting, no editorial fashion grading, no professional beauty filter, no ring-light reflection in the eyes
- Believable mirror physics — the phone is visible in the reflection partially in front of the face, the mirror frame edge appears, background depth is preserved behind the person, no impossible reflection geometry
- No text, no captions, no watermarks, no IG-UI overlays anywhere in the image
- Single image output — one 1:1 photo, no contact sheet, no variant grid, no mirror-vs-original split
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears in the background

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear photo of yourself in good light (this is the face, hair, body, and skin tone the AI will preserve — without it the selfie will not look like you)
- {LOCATION} = a stylish minimal bedroom with soft natural daylight (or pick one — see Bonus tips below for swaps)
- {OUTFIT_STYLE} = a fitted neutral knit top with relaxed jeans (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single 1:1 square hyperrealistic product photograph of a 7-inch collectible action figure version of the exact person from the uploaded reference image, displayed inside a retail blister-pack toy box on a clean off-white retail shelf. The figurine — preserve their face shape, eye shape and color, nose, lip shape, hair color and texture, and skin tone exactly so the toy version is unmistakably them, just rendered as a hand-painted collectible — stands in a neutral confident pose inside a vacuum-formed clear plastic blister against a printed cardboard backer. The cardboard backer has a bold printed header reading {HEADER_LABEL} in clean modern sans-serif type at the top, with small mock collectible-edition graphics — a tiny readable barcode, a small "SERIES 01" sticker in the lower corner, and a small "© 2026" mark — all sized realistically and clearly legible. Around the figurine inside the blister, neatly arranged in their own molded plastic compartments: {MINI_ACCESSORIES}, each rendered to plausible doll-house scale. Soft retail-product-photography lighting, slight depth of field, the box angled at a gentle three-quarter perspective so the buyer can see both the printed cardboard front and the depth of the blister. Render the figurine with believable plastic-toy materiality — matte injection-molded plastic with subtle seam lines along the limbs, hand-painted facial detail with visible micro-brush imperfection, fabric-look outfit textures — not waxy AI-plastic skin. Background of the shelf is clean and uncluttered so the box is the unambiguous focus. Single 1:1 square hyperrealistic product photograph of a personalized collectible figurine in retail packaging, identity-locked to the uploaded person.

**Rules the AI must follow:**
- Aspect ratio 1:1 square — strict, locked at the start and the end of the prompt (works for IG profile crop + Pinterest re-pin + X profile)
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's face shape, eye shape and color, nose, lip shape, hair, and skin tone exactly; the figurine's hand-painted face must be recognizable as the same person, not a generic toy
- Plastic-toy materiality required — matte injection-molded plastic with subtle seam lines, visible hand-paint micro-detail on the face, fabric-look outfit textures, gentle plastic shine on the blister; no waxy "AI plastic" skin on the figurine's face, no porcelain-smooth surfaces
- Render the {HEADER_LABEL} text exactly as specified, in clean modern sans-serif at the top of the cardboard backer, spelled correctly, English Latin script — and render the small printed marks (barcode lines, "SERIES 01" sticker, "© 2026" mark) as legible printed elements; no Lorem Ipsum, no garbled text anywhere
- Single image output — one photograph of one boxed figurine; no contact sheet, no character variant grid, no before/after comparison, no multiple boxes on the shelf
- No watermarks, no extra captions, no logos beyond the specified header and edition marks
- Output the image directly without explaining the prompt back
- All text in English Latin script

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear front-facing photo of yourself in good light (this is the face the AI will hand-paint onto the figurine's head — without it the toy won't look like you)
- {HEADER_LABEL} = JANE — FOUNDER EDITION (or pick one — see Bonus tips below for swaps)
- {MINI_ACCESSORIES} = a tiny laptop, a miniature coffee cup, small headphones, and a stack of mini sticky notes (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single photoreal 4:5 portrait candid selfie of the person in the uploaded reference image, shoulder-up framing, looking slightly off-camera with a relaxed natural expression and a soft genuine closed-mouth smile, wearing a simple {OUTFIT_STYLE} top, a natural no-product-on-skin look (skin looks like real skin, not airbrushed), {LOCATION_VIBE} softly visible behind and slightly out of focus. Smartphone front-camera realism — handheld natural angle, slight imperfect framing, the kind of selfie a real person actually takes. Skin shows visible pores, faint freckles or minor blemishes, and slight micro-asymmetry; this is a real face, not a filter. Identity matches the uploaded reference exactly — same bone structure, same eyes, same nose, same lips, same proportions. Single 4:5 portrait photo, candid, intimate, real-phone feel.

**Rules the AI must follow:**
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's face exactly (bone structure, eyes, nose, lips, skin tone, proportions); the person in the photo must be unambiguously recognizable as the person in the reference
- Realistic skin texture required — visible pores, faint freckles, slight micro-asymmetry; no porcelain skin, no over-retouched look, no AI plastic smoothing, no beauty filter
- Smartphone front-camera realism — handheld natural angle, slight imperfect framing OK; no DSLR studio editorial lighting, no professional beauty look, no magazine grading
- Single image output — one 4:5 photo, no contact sheet, no variant grid, no before/after split
- No text, captions, watermarks, or logos anywhere in the image
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears in the background

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, front-facing portrait of yourself (this is the face the AI will keep — without it the photo will not look like you)
- {OUTFIT_STYLE} = casual minimalist neutral-color top (or pick one — see Bonus tips below)
- {LOCATION_VIBE} = a clean lived-in living room with soft natural daylight from a window (or pick one — see Bonus tips below)

**Generate this image:**

A single hyper-realistic vertical 4:5 photograph of the person in the uploaded reference image, sitting deeply relaxed in a comfortable living room setting — {HOME_VIBE} — naturally engaged in {ACTIVITY}, not looking at the camera, completely absorbed in the moment. Camera angle is a high side view from a doorway, as if a friend wandered by and quietly took the photo. Soft warm afternoon light through a large window, warm reflected light bouncing off the walls, no harsh shadows. Shot on a Sony A7R V, 50mm f/1.4 lens, shallow depth of field, natural skin and fabric textures, photoreal smartphone-candid feel rather than editorial. Identity matches the uploaded reference photo exactly — same bone structure, same eyes, same nose, same lips, same proportions. Single 4:5 portrait photo, candid lifestyle, real-home feel.

**Rules the AI must follow:**
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's face exactly (bone structure, eyes, nose, lips, skin tone, proportions); the person must be unambiguously recognizable as the person in the reference
- Realistic skin and fabric texture required — visible pores, faint freckles, slight micro-asymmetry, natural fabric weave; no plastic skin, no CGI look, no AI artifacts, no beauty filter, no over-retouching
- The person is naturally engaged in the activity — eyes on what they are doing, not on the camera; this is a candid, not a posed portrait
- Single image output — one 4:5 photo, no contact sheet, no variant grid, no multiple-angle composite
- No text, captions, watermarks, or logos anywhere in the image
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears in the background

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, front-facing portrait of yourself (this is the face the AI will keep — without it the photo will not look like you)
- {HOME_VIBE} = a comfortable living room with a soft throw blanket on the sofa, a houseplant in the corner, and a window with warm afternoon light (or pick one — see Bonus tips below)
- {ACTIVITY} = reading an open book on their lap with a warm mug of coffee on a side table (or pick one — see Bonus tips below)

**Generate this image:**

A single photoreal 4:5 portrait candid travel photo of the person in the uploaded reference image, {ACTIVITY_SCENE}, wearing {TRAVEL_OUTFIT}, natural relaxed posture, looking off into the distance with a soft genuine expression — not posing for the camera, just in the moment. Soft warm golden-hour light from the side, slight motion blur on the surroundings (suggesting a real handheld capture rather than a posed studio shot), shallow depth of field, shot on a 50mm f/1.8 lens. Photoreal smartphone-candid feel, the kind of photo a friend would take, not an editorial fashion shoot. Identity matches the uploaded reference photo exactly — same bone structure, same eyes, same nose, same lips, same proportions. Single 4:5 portrait photo, candid outdoor lifestyle, real-travel feel.

**Rules the AI must follow:**
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's face exactly (bone structure, eyes, nose, lips, skin tone, proportions); the person must be unambiguously recognizable as the person in the reference
- Realistic skin texture required — visible pores, faint freckles, slight micro-asymmetry; no porcelain skin, no over-retouched look, no AI plastic smoothing, no beauty filter, no glossy magazine grading
- Candid lifestyle aesthetic — looks like a friend captured the moment, not a fashion editorial; relaxed natural posture, not a posed model stance
- Single image output — one 4:5 photo, no contact sheet, no variant grid, no multiple-angle composite
- No text, captions, watermarks, or logos anywhere in the image
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears in the background

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, front-facing portrait of yourself (this is the face the AI will keep — without it the photo will not look like you)
- {ACTIVITY_SCENE} = walking along a coastal cliff trail with the ocean visible behind and soft sea-mist in the distance (or pick one — see Bonus tips below)
- {TRAVEL_OUTFIT} = comfortable weekend layers — fitted t-shirt with a light jacket and casual pants in neutral earth tones (or pick one — see Bonus tips below)

**Generate this image:**

A single photoreal 4:5 portrait candid photo of the person in the uploaded reference image, genuinely absorbed in {HOBBY} — caught mid-action, eyes on the activity (not on the camera), natural absorbed expression. Wearing {HOBBY_OUTFIT}, in the indoor or outdoor setting that naturally matches the hobby, with soft natural daylight (window light if indoor, soft daylight if outdoor), shallow depth of field, slight smartphone-candid feel as if a friend captured the moment. Identity matches the uploaded reference photo exactly — same bone structure, same eyes, same nose, same lips, same proportions. Realistic skin texture, natural fabric texture, real-world micro-imperfection. Single 4:5 portrait photo, candid hobby-in-action, real-life feel.

**Rules the AI must follow:**
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's face exactly (bone structure, eyes, nose, lips, skin tone, proportions); the person must be unambiguously recognizable as the person in the reference
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt
- Eyes on the activity, not on the camera — this is a candid mid-action shot, not a posed portrait holding a prop
- Realistic skin texture required — visible pores, faint freckles, slight micro-asymmetry; no porcelain skin, no over-retouched look, no AI plastic smoothing, no beauty filter, no editorial magazine grading
- Single image output — one 4:5 photo, no contact sheet, no variant grid, no multi-panel collage, no before/after split
- No text, captions, watermarks, or logos anywhere in the image
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears in the background

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, front-facing portrait of yourself (this is the face the AI will keep — without it the photo will not look like you)
- {HOBBY} = playing acoustic guitar mid-strum on a sunlit couch with eyes on the strings (or pick one — see Bonus tips below)
- {HOBBY_OUTFIT} = a soft fitted t-shirt and casual relaxed pants in a neutral tone (or pick one to match the hobby — see Bonus tips below)

**Generate this image:**

A single photoreal 2:3 vertical Pinterest pin, 1000×1500 ratio, designed for maximum thumbnail readability. Top 45% of the pin is a soft-focused portrait of {PORTRAIT_SUBJECT}, eyes calm, looking slightly off-camera, with a gentle gradient fading the bottom edge of the portrait into the quote area below. Bottom 55% of the pin is a clean quote area on a {BACKGROUND_TONE} background. Centered in the quote area is the exact quote text "{QUOTE_TEXT}" set in a large elegant light-gold serif typeface, rendered exactly as written with correct spelling and punctuation. Below the quote, on its own line, the attribution "— {ATTRIBUTION}" appears in the same serif in a smaller size. A large subtle quotation-mark glyph sits behind the first line of the quote at low opacity as a graphic accent. Generous breathing room around the text — text occupies roughly 70% of the quote area width, with comfortable margins. Soft natural light on the portrait, calm and contemplative mood, magazine-editorial Pinterest aesthetic, single 2:3 vertical pin.

**Rules the AI must follow:**
- Aspect ratio 2:3 vertical (1000×1500) — strict, locked at the start and the end of the prompt
- Render the quote text "{QUOTE_TEXT}" and attribution "— {ATTRIBUTION}" exactly as written, with correct English spelling and punctuation — no Lorem Ipsum, no AI text gibberish, no rewriting, no extra words
- Text must remain readable when the pin is shown as a small mobile Pinterest thumbnail — large headline-grade serif, high contrast against the background tone
- Single image output — no contact sheet, no two-pin variant grid, no before/after split
- Realistic portrait texture — visible skin texture, natural light, no over-retouched plastic skin, no AI plastic look
- No watermarks, no website URLs, no logos other than what is specified above
- All text in English Latin script — no substitution to other writing systems
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- {QUOTE_TEXT} = The quieter you become, the more you can hear
- {ATTRIBUTION} = Ram Dass
- {PORTRAIT_SUBJECT} = a calm middle-aged man with kind eyes and silver hair, soft natural daylight on his face (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single 2:3 vertical Pinterest pin, 1000×1500 ratio, built as a dramatic perspective-typography cover for a tutorial or blog post. The headline phrase "{HEADLINE_TEXT}" is rendered exactly as written in heavy, condensed grotesque sans-serif (Helvetica Black Condensed / Akzidenz-Grotesk Bold feel), set in {ACCENT_COLOR_TEXT} on a {BACKGROUND_COLOR} background. The headline fills roughly 70-80% of the pin width and is the unambiguous hero of the composition, positioned in the upper-middle band of the pin. Each letter of the headline is duplicated in multiple receding perspectives converging to a vanishing point behind the text, with repeated thin {BACKGROUND_COLOR}-contrasting outlines creating depth and motion illusion — isometric-inspired but warped, bold graphic-design energy. A short single-line subheading "{SUBHEADING_TEXT}" sits below the headline in a much smaller clean sans-serif, also rendered exactly as written. Generous breathing room around all text. The pin reads instantly at thumbnail size on a mobile Pinterest feed. Bold graphic-design poster aesthetic, single flat 2:3 vertical pin.

**Rules the AI must follow:**
- Aspect ratio 2:3 vertical (1000×1500) — strict, locked at the start and the end of the prompt
- Render "{HEADLINE_TEXT}" and "{SUBHEADING_TEXT}" exactly as written, with correct English spelling — no Lorem Ipsum, no AI gibberish, no rewriting, no extra words
- Headline text is the single hero — must dominate the composition and be instantly readable at small mobile thumbnail size
- Single image output — no contact sheet, no variant grid, no multi-pin layout
- High contrast between text color and background — Pinterest algorithm rewards readability on small thumbnails
- No watermarks, no website URLs, no logos other than what is specified above
- All text in English Latin script — no substitution to other writing systems
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- {HEADLINE_TEXT} = 5 Ways To Save More
- {SUBHEADING_TEXT} = Money habits that actually stick
- {ACCENT_COLOR_TEXT} = bright acid yellow (background is matte charcoal gray — or pick a different palette in Bonus tips below)

**Generate this image:**

A single photoreal 2:3 vertical Pinterest pin, 1000×1500 ratio, top-down flat-lay photograph of a {THEME_SCENE} on a warm rustic wooden tabletop. Five clearly visible objects arranged in a relaxed asymmetric composition with comfortable breathing room around them, each object positioned with a clear margin from the others so the layout reads at small thumbnail size. Soft natural daylight from the upper-left, warm cozy tones, lifestyle-photography quality. Over the photograph (not replacing it) are white hand-drawn doodle outlines and arrows around each object, drawn as if with a fine white marker, with playful handwritten-style labels — exactly these five labels, rendered exactly as written and spelled correctly, each label pointing at the object it describes: "{LABEL_1}", "{LABEL_2}", "{LABEL_3}", "{LABEL_4}", "{LABEL_5}". Labels are short, single English phrases, set in a casual handwritten font, all five visually consistent in style and stroke width. At the very top of the pin, in a slightly larger handwritten style in the same white, the title text "{PIN_HEADLINE}" rendered exactly as written. Warm cozy Instagram-flat-lay aesthetic, single 2:3 vertical Pinterest pin.

**Rules the AI must follow:**
- Aspect ratio 2:3 vertical (1000×1500) — strict, locked at the start and the end of the prompt
- Render every label and the headline exactly as written above with correct English spelling — no Lorem Ipsum, no AI text gibberish, no rewriting, no extra labels added, no labels removed
- Exactly five objects in the flat lay, each with one doodle label and one arrow — no extra objects floating in, no objects without labels
- Single image output — no contact sheet, no variant grid, no before/after split
- Realistic flat-lay photography texture — visible wood grain, natural light, real object materials; no plastic / waxy / over-retouched look
- The doodle layer is hand-drawn-style only — playful imperfect strokes, not vector-clean lines
- All text in English Latin script — no substitution to other writing systems
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- {THEME_SCENE} = a slow Sunday morning routine, with a ceramic mug of coffee, a glazed donut, a small succulent in a white pot, a vintage hardcover book, and round eyeglasses
- {PIN_HEADLINE} = my slow Sunday routine
- {LABEL_1} = morning brew · {LABEL_2} = sweet treat · {LABEL_3} = little green friend · {LABEL_4} = current read · {LABEL_5} = before contacts

**Generate this image:**

A single 2:3 vertical Pinterest pin, 1000×1500 ratio, laid out as an 8-photo cinematic storyboard collage of one continuous day in the life of the same person from the uploaded reference image, arranged in a clean 2-column × 4-row grid with thin consistent white gutters between every frame and no outer border. All 8 frames feature the same person from the reference photo with strict identity consistency — same bone structure, same eyes, same nose, same lips, same proportions, same skin tone — across every single frame. The 8 frames tell a {THEME_NARRATIVE} narrative arc in order, top-to-bottom, left-to-right:

- Frame 1: morning wake-up scene, soft window light, the person stretching or yawning in a calm at-home outfit
- Frame 2: getting ready, mirror or window reflection, the person putting on a casual outfit appropriate to the day theme
- Frame 3: leaving home, the person stepping out the door, soft morning street light behind
- Frame 4: in-transit moment, the person on the way to the day's main activity, candid mid-action
- Frame 5: arrival / the main activity beginning, the person in the day's primary setting
- Frame 6: a peak moment of the day, the person in their most expressive expression of the narrative
- Frame 7: a quiet downbeat — the person resting, eating, or pausing mid-day
- Frame 8: golden-hour wrap, the person heading home or sitting still at the end of the day, warm sunset light

Each frame is a photoreal smartphone-quality image, cinematic editorial color grading, warm vibrant tones, single coordinated palette across all 8 frames. The pin reads as one composition at thumbnail size — frames are clearly distinct but visually unified. Single 2:3 vertical Pinterest pin, 8-frame storyboard layout, identity-locked across all frames.

**Rules the AI must follow:**
- Aspect ratio 2:3 vertical (1000×1500) — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's bone structure, eyes, nose, lips, proportions, and skin tone exactly; the same person must be unmistakably recognizable in all 8 frames
- Exactly 8 frames in a 2-column × 4-row grid layout with thin consistent white gutters — no extra frames, no missing frames, no irregular layout
- Each frame shows a clearly different scene, mood, and time of day per the narrative arc above — no duplicate frames, no repeated angles
- Realistic skin texture in every frame — visible pores, natural micro-asymmetry; no over-retouched plastic skin in any frame
- Single image output — one 2:3 collage pin; do not output 8 separate images
- No text, captions, frame numbers, or watermarks anywhere in any of the 8 frames or in the gutters
- All text in English Latin script if any incidental signage appears in a background
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, front-facing photo of yourself (this is the face the AI will keep consistent across all 8 frames — without it the person will drift across panels and not look like you)
- {THEME_NARRATIVE} = a cozy slow Sunday at home (or pick one — see Bonus tips below for swaps: shopping day / café date / gym day / wedding day / travel day)

**Generate this image:**

A single photoreal 16:9 horizontal YouTube thumbnail of a podcast host facing the camera with a confident, warm, slightly smiling expression. The host's face must match the uploaded reference photo exactly — same bone structure, eyes, nose, lips, hair, and skin tone — and is positioned on the LEFT third of the frame, occupying roughly 30-50% of the frame width, shoulders-up to chest-up framing. The host wears a smart-casual outfit such as a neutral knit or button-up, sitting in a comfortable studio chair. A high-quality podcast microphone on a boom arm is visible close to their face on the left. The studio background is softly defocused with a clean cinematic atmosphere and a warm-to-cool color gradient accented by {ACCENT_COLOR} rim lighting; subtle props like a notebook and audio interface sit softly out of focus on the desk. The RIGHT two-thirds of the frame holds the bold high-contrast headline text "{HEADLINE_TEXT}" in a large clean sans-serif, all letters spelled exactly as typed, no Lorem Ipsum, no hallucinated text, no extra words; the headline is the second-most-prominent element after the face. Soft key lighting on the face, subtle rim lighting from the accent color, shallow depth of field, ultra-realistic skin texture with visible pores and slight micro-asymmetry, 50mm lens look, f/1.8, professional color grading, high contrast. Single 16:9 horizontal YouTube podcast thumbnail, face left, headline right.

**Rules the AI must follow:**
- Aspect ratio 16:9 horizontal — strict, locked at the start and the end of the prompt (1280×720 YouTube thumbnail format)
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's face exactly (bone structure, eyes, nose, lips, skin tone); do not stylize, retouch, or "improve" the face
- Render the headline text "{HEADLINE_TEXT}" exactly as typed, every letter spelled correctly — no Lorem Ipsum, no AI-hallucinated text, no extra words, no auto-translation to other writing systems
- Face occupies the LEFT third of the frame at 30-50% of frame width; headline text occupies the RIGHT two-thirds at 40-50% of frame width — left-to-right scan path
- Realistic skin texture required — visible pores, slight micro-asymmetry, real studio lighting unevenness; no porcelain skin, no over-retouched look, no AI plastic
- Single image output — no contact sheet, no variant grid, no multi-panel split, no before/after layout
- Output the image directly without explaining the prompt back
- All text in English Latin script

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit front-facing photo of yourself (this is the face the AI will keep identical — without it the thumbnail will not look like you)
- {HEADLINE_TEXT} = HOW I GREW TO 100K (or pick your own — keep it 3-5 words max, all caps reads best at thumbnail size)
- {ACCENT_COLOR} = warm amber and soft teal (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single photoreal 16:9 horizontal YouTube thumbnail of a content creator sitting at a desk in a modern creator studio. The creator's face must match the uploaded reference photo exactly — same bone structure, eyes, nose, lips, hair, and skin tone — with a confident slight smile, eyes meeting the camera, positioned on the RIGHT third of the frame occupying roughly 30-50% of the frame width, chest-up framing. In front of the creator is a mounted boom-arm microphone with a studio microphone. The creator wears a comfortable creator outfit such as a neutral hoodie or jacket. The studio walls show subtle acoustic foam panels for a soundproof studio look, and {LED_COLOR} LED tube lights glow in the background for a cinematic YouTube aesthetic. The LEFT two-thirds of the frame holds the bold high-contrast headline text "{HEADLINE_TEXT}" in a large clean sans-serif, all letters spelled exactly as typed, no Lorem Ipsum, no hallucinated text, no extra words; the headline is the second-most-prominent element after the face. Soft professional key lighting on the face plus subtle rim lighting from the LED accents, ultra-realistic skin texture with visible pores and slight micro-asymmetry, sharp focus on creator + headline, slightly soft focus on background props. Vibrant but clean colors, modern minimal feel. Single 16:9 horizontal YouTube creator-studio thumbnail, face right, headline left.

**Rules the AI must follow:**
- Aspect ratio 16:9 horizontal — strict, locked at the start and the end of the prompt (1280×720 YouTube thumbnail format)
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's face exactly; do not stylize, "improve", or substitute a different face
- Render the headline text "{HEADLINE_TEXT}" exactly as typed, every letter spelled correctly — no Lorem Ipsum, no AI-hallucinated text, no extra words, no auto-translation
- Face occupies the RIGHT third at 30-50% of frame width; headline occupies the LEFT two-thirds at 40-50% of frame width
- Realistic skin texture required — visible pores, slight micro-asymmetry, real lighting unevenness; no porcelain skin, no over-retouched look, no AI plastic
- Single image output — no contact sheet, no variant grid, no multi-panel split
- Output the image directly without explaining the prompt back
- All text in English Latin script

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit front-facing photo of yourself (this is the face the AI will keep identical — without it the thumbnail will not look like you)
- {HEADLINE_TEXT} = MY FIRST 30 DAYS (or pick your own — keep it 3-5 words max, all caps reads best at thumbnail size)
- {LED_COLOR} = blue and pink (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single 16:9 horizontal hand-drawn illustrated YouTube thumbnail in a simple, clean, editorial illustration style. The thumbnail features an illustrated portrait of the person in the uploaded reference photo, recognizable as the same person but rendered in hand-drawn style — preserve their distinctive features (hair shape, glasses if any, skin tone, gender presentation, general age) while translating from photo to illustration. The person is positioned on the LEFT side of the frame, shoulders-up, with a friendly engaged expression, occupying roughly 35-45% of the frame width. The RIGHT side of the frame holds the bold high-contrast headline text "{HEADLINE_TEXT}" rendered in a clean modern sans-serif (not a hand-drawn font — the headline must be readable at thumbnail size), all letters spelled exactly as typed, no Lorem Ipsum, no hallucinated text, no extra words. The background uses a soft {COLOR_PALETTE} gradient with simple hand-drawn doodle accents (small lines, dots, simple shapes) scattered lightly around the person — editorial illustration feel, not cluttered. Visible hand-drawn line work with slight imperfection (organic line weight variation, intentional micro-roughness), flat color fills with light texture, no photoreal rendering, no 3D. Single 16:9 horizontal hand-drawn illustrated thumbnail, illustrated portrait left, headline right.

**Rules the AI must follow:**
- Aspect ratio 16:9 horizontal — strict, locked at the start and the end of the prompt (1280×720 YouTube thumbnail format)
- Render the headline text "{HEADLINE_TEXT}" exactly as typed, every letter spelled correctly — no Lorem Ipsum, no AI-hallucinated text, no extra words, no auto-translation to other writing systems
- The illustrated portrait must be recognizable as the same person from the uploaded reference photo — keep their hair, glasses, skin tone, gender presentation, and general age; this is illustrated likeness, not photoreal identity-lock
- Hand-drawn illustration style only — visible line work, slight organic imperfection, flat color fills with light texture; no photoreal rendering, no 3D rendering, no AI plastic, no glossy CGI look
- Single image output — no contact sheet, no variant grid, no multi-panel split
- Output the image directly without explaining the prompt back
- All text in English Latin script

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit photo of yourself (the AI will translate your face into illustration style — without it the illustrated portrait will not look like you)
- {HEADLINE_TEXT} = HOW AI ACTUALLY WORKS (or pick your own — keep it 3-5 words max, all caps reads best at thumbnail size)
- {COLOR_PALETTE} = soft blue and warm coral (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single 16:9 horizontal 3D Pixar-style hook thumbnail for an explainer video. The character is a Pixar-style 3D-rendered version of the person in the uploaded reference photo — recognizable as the same person (preserve their hair, skin tone, glasses if any, gender presentation, and general age) but translated into stylized 3D animation style. The character has a surprised, realizing, pointing-at-something expression, positioned on the RIGHT side of the frame, occupying roughly 25-35% of the frame width, with a dynamic mid-action pose. The scene visualizes the topic "{VIDEO_TOPIC}" as a clear before-vs-after transformation in the background: the old / broken / outdated version on one side glitching or breaking down, the new / improved / transformed version glowing and pulsing with vibrant energy on the other side, with the character pointing toward the transformation. A subtle foreground element (a relevant tool, gadget, or prop softly blurred) sits in the lower-left corner to create depth. The LEFT side of the frame holds the bold high-contrast headline text "{HEADLINE_TEXT}" rendered in bold white with a soft glow, taking 70% of visual weight (the headline dominates the composition; the character takes 30%), all letters spelled exactly as typed, no Lorem Ipsum, no hallucinated text, no extra words. Cinematic lighting with one bright neon accent color highlighting the transformation, vibrant saturated Pixar-bright palette, depth-of-field with blurred foreground prop → sharp character and headline → slightly soft background. Single 16:9 horizontal 3D Pixar-style YouTube thumbnail, text-dominant 70/30 composition, headline left, character right.

**Rules the AI must follow:**
- Aspect ratio 16:9 horizontal — strict, locked at the start and the end of the prompt (1280×720 YouTube thumbnail format)
- Render the headline text "{HEADLINE_TEXT}" exactly as typed, every letter spelled correctly — no Lorem Ipsum, no AI-hallucinated text, no extra words, no auto-translation to other writing systems
- The 3D character must be recognizable as the person from the uploaded reference photo — keep their hair, skin tone, glasses, gender presentation, and general age; this is stylized likeness, not photoreal identity-lock
- 70/30 text-dominant composition — the headline occupies roughly 70% of visual weight on the left, the character occupies roughly 30% on the right
- 3D Pixar-style rendering only — clean stylized 3D, vibrant Pixar-bright color, cinematic lighting; no photoreal rendering, no flat 2D, no AI plastic surface
- Single image output — no contact sheet, no variant grid, no multi-panel split
- Output the image directly without explaining the prompt back
- All text in English Latin script

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit photo of yourself (the AI will translate your face into Pixar-style 3D — without it the character will not look like you)
- {HEADLINE_TEXT} = THE BIG SHIFT (or pick your own — keep it 2-4 words max, bold short hooks read best at thumbnail size)
- {VIDEO_TOPIC} = old school methods becoming AI-powered methods (or describe your video's before-vs-after transformation in one short phrase — see Bonus tips below for examples)

**Generate this image:**

A single photoreal 4:5 portrait main listing photo for an Etsy / Shopify product page, product-only with no person in frame, showing the exact product from the uploaded reference image — match its shape, color, material, label text, dimensions, and surface details exactly. The product is positioned dead-center and occupies roughly 55-65% of the frame, shot at eye level so the product looks true-to-life and accurate (not exaggerated by a low or high angle). The overall mood is {BRAND_MOOD}, and the product sits on {BACKGROUND_SURFACE} with soft natural daylight from the upper-left, casting a gentle realistic shadow to the lower-right. Background is softly defocused but stays simple and uncluttered so the product remains the unambiguous hero. Catalog-grade clarity — every material, texture, and edge of the product matches the uploaded reference and is sharp and legible. Single 4:5 portrait main listing image, product-only, eye-level, hero-centered.

**Rules the AI must follow:**
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt
- Product fidelity is the highest-priority constraint — match the uploaded reference photo's shape, color, material, label text, and surface details exactly; do not invent or substitute a different product
- The product itself is the single hero — no person, no hands, no model in the frame
- Product anchored dead-center, occupying 55-65% of the frame, with clean breathing room on all sides
- Eye-level camera angle — no exaggerated low-angle hero perspective, no top-down flat lay (this is a true-to-product catalog shot, not a campaign poster)
- Realistic material texture required — preserve the actual texture of the uploaded product (visible weave / grain / brush marks / surface micro-imperfection); no over-polished glassy "AI plastic" surface, no waxy highlights
- No new on-image text added, no logos added, no captions, no watermarks, no price tags; preserve any text that exists on the actual product in the reference photo
- Single image output — no contact sheet, no variant grid, no before/after split, no multiple angles in one frame
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit photo of your product (this is what the AI will keep faithful — shape, color, material, label text, and surface details all preserved)
- {BRAND_MOOD} = clean minimal premium with a calm, modern, slightly editorial feel (or pick one — see Bonus tips below for swaps)
- {BACKGROUND_SURFACE} = a smooth seamless off-white paper backdrop curving gently into the floor (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A clean lifestyle product photograph for an Etsy listing, 4:5 portrait aspect, showing a single human hand holding {PRODUCT_DESCRIPTOR} matching the uploaded reference image exactly, hand entering the frame from the lower right, product centered slightly above the midline at the visual hero spot, hand and product in sharp focus, set against a {SETTING_VIBE} background with gentle soft shadows, warm natural daylight from a window on the left, soft falloff toward the edges, shallow depth of field that gently blurs the background while keeping the product crisp, a single complementary prop just visible in the soft-focus background to suggest context without competing for attention, neutral cream-and-linen color palette, calm and inviting Etsy small-shop mood — not a glossy advertising campaign — composed with comfortable negative space around the product so the listing thumbnail still reads at small sizes, photoreal smartphone-quality realism, 4:5 portrait aspect.

**Rules the AI must follow:**
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt
- Product fidelity is the highest-priority constraint — match the uploaded reference photo's shape, proportions, material, color, label, and any text on the product exactly; do not invent a different product
- Show only one hand, no face, no full body, no second person; the hand should look natural and gender-neutral (relaxed fingers, no manicure styling, no jewelry beyond a plain band at most)
- Realistic imperfection required — visible skin texture and pores on the hand, faint natural creases, subtle surface texture on the background, gentle imperfect lighting; no porcelain skin, no plastic AI smoothing, no over-retouched gloss
- Render any text that appears on the product or its label exactly as it is on the uploaded reference — correctly spelled, no Lorem Ipsum, no garbled letters, no extra invented logos
- Single image output — one 4:5 photo, not a moodboard, not a contact sheet, not a variant grid
- All text in English Latin script — no substitution to other writing systems
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit reference photo of your actual product (front view + any side with a label visible) — without this the AI will invent a generic product that won't match your listing
- {PRODUCT_DESCRIPTOR} = an 8 oz amber soy candle in a frosted glass jar with a minimalist kraft label
- {SETTING_VIBE} = sunlit linen-covered table next to a window, with a small sprig of dried lavender out of focus in the background

**Generate this image:**

Create a 1:1 square close-up detail product photograph of the {PRODUCT_TYPE} from the uploaded reference image, lying diagonally on a smooth {SURFACE_MATERIAL} surface in hard late-afternoon natural sunlight from the upper right. The product occupies roughly 55% of the frame, off-center to the lower-left third, with the camera angled slightly above to reveal both the top surface and one side edge so the buyer can read the material, texture, and craftsmanship up close. Strong directional sunlight casts one crisp architectural shadow stretching diagonally from the product across the surface, with a softer secondary shadow trailing behind, creating a quiet editorial mood without distracting from the product. The visible label or marking on the product reads exactly "{ON_PRODUCT_TEXT}" in clean minimal typography matching the reference photo, spelled exactly as written with no extra characters. The surface shows authentic natural texture — fine grain, subtle imperfections, faint shadow gradient — and the background is the same surface continuing softly out of focus, no props, no styling clutter. Output a single 1:1 square image with magazine-quality detail clarity on the product and the same surface material continuing to the edges of the frame.

**Rules the AI must follow:**
- Aspect ratio 1:1 square — strict, locked at the start and the end of the generation
- No Lorem Ipsum / no garbled text — render the label exactly as specified in {ON_PRODUCT_TEXT}, spelled correctly, in the same script and casing
- Single image output — one clean detail shot, no contact sheet, no variant grid, no before/after split
- Realistic imperfection required — visible material grain, fine surface variation, subtle dust or fiber if natural to the material; no flat AI-plastic look, no over-retouched glossy sheen
- Output the image directly without explaining the prompt back
- All text in English Latin script — no substitution to other writing systems
- Product identity lock — match the uploaded reference photo's exact shape, proportions, color, and label placement; this is a real product the seller is photographing, not a stylized reinterpretation
- Honest commercial detail shot — sharp focus on product material and craftsmanship, no decorative props, no flower petals, no scattered confetti, no styled lifestyle clutter

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit photo of your actual product (the AI uses this to lock the real shape, color, and label of what you're selling — without it, the photo will not look like your product)
- {PRODUCT_TYPE} = handmade lavender soap bar (or: ceramic mug / leather card holder / brass candle / linen tea towel / jewelry pendant / wax-sealed envelope / your product type)
- {SURFACE_MATERIAL} = warm beige limewashed plaster (or: pale concrete / natural oak wood / cream linen / black slate / unbleached paper / brushed travertine)
- {ON_PRODUCT_TEXT} = lavender (or your exact brand name / product label / "no text" if your product has no visible text)

**Generate this image:**

A single 1:1 square product photograph of the exact product shown in the uploaded reference image, placed centered on a clean matte-white circular podium, with a {SCALE_REFERENCE} sitting beside the product at the same eye level on the same podium surface to communicate real-world size. The background is a soft seamless gradient from light cream at the top to pure white at the bottom, no other props, no text frame, no shadows fighting for attention. Lighting is soft diffused studio light from the upper left, premium Apple-keynote feel — even, gentle, no harsh shadows, with one subtle contact shadow under the podium so the product looks grounded. Below the podium, centered, a single line of small clean sans-serif type reads exactly "{DIMENSION_TEXT}" — render that text exactly as written, no rewording, no extra words. The {PRODUCT_NAME} must remain unmistakably the same product as the uploaded reference — same shape, same color, same material, same surface details, same proportions. Composition: product slightly left-of-center, scale reference slightly right, both fully visible, minimum 30% empty space around the pair, 1:1 square aspect, photoreal product photography.

**Rules the AI must follow:**
- Aspect ratio 1:1 square — strict, locked at start and end, suitable for Etsy listing thumbnail
- Product fidelity is the highest-priority constraint — match the uploaded reference photo's shape, color, material, surface texture, and proportions exactly; do not stylize, redesign, or "improve" the product
- Render the dimension text exactly as written inside the prompt — no Lorem Ipsum, no rewording, no garbled letters, correct spelling, sans-serif, small, centered below the podium
- The scale reference object must sit on the same podium surface at the same height as the product, not floating, not behind, not in front — side by side so size comparison is immediate
- Single image output — no contact sheet, no variant grid, no before/after split, no multiple angles
- Realistic product photography imperfection — natural micro-texture on the podium surface, subtle real lighting falloff, no over-rendered CGI plastic look
- Output the image directly without explaining the prompt back
- All text in English Latin script — no substitution to other writing systems

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit photo of the actual product you sell (front or three-quarter angle, plain background ideal — this is the product the AI will keep identical in the scale shot)
- {PRODUCT_NAME} = ceramic pour-over coffee dripper (replace with your product, e.g. leather card wallet / silver hoop earrings / scented soy candle 8oz)
- {SCALE_REFERENCE} = a US quarter coin (or pick one: a standard #2 pencil / an iPhone 15 / an AA battery / an open adult hand palm-up / a credit card)
- {DIMENSION_TEXT} = 4.2 in tall · 5.1 in wide (write your real dimensions exactly as you want them rendered, e.g. "3.5 cm × 2 cm" or "fits in the palm of your hand")

**Generate this image:**

A single photoreal 1:1 square top-down flat-lay product mockup of a neatly folded crew-neck cotton t-shirt placed on {BACKGROUND_SURFACE}, shot directly overhead. The t-shirt is {TEE_COLOR} with soft realistic fabric texture (visible cotton weave, gentle fold creases, slight shading where the fabric folds under itself). The chest print area shows the design from the uploaded reference image — apply it exactly as uploaded, preserving the original colors, line weights, typography, and proportions of the design; the print sits centered on the visible chest panel and partially clips at the natural fold line where the bottom edge of the shirt folds under, for true flat-lay realism. The print is rendered as if screen-printed or DTG-printed onto the fabric — slightly absorbed into the cotton weave (not floating like a sticker, not glassy), with subtle texture interaction so the print follows the fabric's micro-folds. Soft diffused studio daylight from the upper-left, very gentle realistic shadow under the folded edges. Background is the surface only, minimal and uncluttered — the folded tee is the unambiguous hero. Catalog-grade clarity: every fiber of the fabric and every detail of the uploaded design is sharp and faithful. Single 1:1 square top-down flat-lay mockup, product-only, no person, no hands.

**Rules the AI must follow:**
- Aspect ratio 1:1 square — strict, locked at the start and the end of the prompt
- Design fidelity is the highest-priority constraint — match the uploaded design's colors, line weights, typography, and proportions exactly; do not redraw, restyle, recolor, or "AI-interpret" the design
- The print must not be tinted by the tee color — preserve the uploaded design's original colors even on a colored tee (POD buyers care about this — washed-out mockups lose sales)
- Print interaction with fabric is required — design slightly absorbed into the cotton weave, following micro-folds; no floating-sticker look, no glossy decal sheen
- Realistic fabric texture — visible cotton weave, gentle fold creases, soft fiber detail; no over-polished waxy plastic surface
- Top-down flat-lay camera only — no angled three-quarter view, no perspective tilt (this is a catalog mockup, not a styled lifestyle shot)
- Single image output — no contact sheet, no front/back split, no multi-color grid
- No added text, no added logos, no captions, no watermarks, no price stickers (the only on-image content is the uploaded design itself)
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: your design artwork (PNG / JPG of the print you're selling — the AI will apply it to the chest panel exactly as uploaded, preserving colors and detail)
- {TEE_COLOR} = soft heather grey with subtle melange texture (or pick one — see Bonus tips below for swaps)
- {BACKGROUND_SURFACE} = a softly grained light oak wood surface with subtle natural texture (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single photoreal 4:5 portrait editorial-macro product mockup of {MUG_STYLE} resting on {SURFACE}, shot side-on at the height of the mug's body so the curved side faces the camera directly. The design from the uploaded reference image is printed on the mug's curved side facing the camera — apply it exactly as uploaded, preserving the original colors, line weights, typography, and proportions, and wrap it realistically around the cylindrical surface so that straight lines in the design follow the gentle curvature of the mug (slightly bending toward the edges, sharp and undistorted in the center facing the lens). The print sits squarely in the center of the visible mug panel, occupying the natural print area and not running off the top rim or the bottom base. Soft visible steam rises gently from the rim, lit from the side so the steam catches a faint warm edge of light against a slightly darker background — classic editorial-macro mood, café merch aesthetic. Strong directional side lighting from the right creates a subtle highlight along the mug's left edge and a soft gradient shadow along the right; the background is uncluttered and slightly out of focus. The ceramic glaze is realistic — soft sheen, not glassy plastic, with very fine surface micro-imperfection. Catalog-grade macro clarity: every detail of the uploaded design and every fiber/grain of the surface is sharp and faithful. Single 4:5 portrait editorial-macro mockup, product-only, no person, no hands.

**Rules the AI must follow:**
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt
- Design fidelity is the highest-priority constraint — match the uploaded design's colors, line weights, typography, and proportions exactly; do not redraw, restyle, or AI-interpret the design
- The print must follow realistic cylindrical curvature — straight lines bend slightly toward the mug's curved edges; the design does not appear pasted flat like a sticker
- Print colors must remain true to the uploaded design — do not let the mug's glossy white reflect a color cast into the print (buyers reject washed-out or tinted mockups)
- Realistic ceramic glaze — soft sheen with very fine surface micro-imperfection; no glassy waxy plastic surface, no over-polished AI gloss
- Side-on camera at mug-body height — no top-down angle, no exaggerated low-angle hero perspective (this is a catalog macro, not an advertising poster)
- Single image output — no contact sheet, no front/back grid, no multi-angle composite
- No added text, no added logos, no captions, no watermarks (the only on-image content is the uploaded design)
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: your design artwork (PNG / JPG of the print you're selling — the AI will wrap it around the mug's curved side facing the camera, preserving colors and detail)
- {MUG_STYLE} = a classic 11oz white ceramic mug with a standard C-handle on the right side of frame (or pick one — see Bonus tips below for swaps)
- {SURFACE} = a dark natural linen tablecloth with soft visible weave (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single 4:5 portrait sticker-sheet design featuring the same character from the uploaded reference image, shown chest-up across six puffy-sticker variants with different facial expressions: neutral, smiling, angry, sad, happy / excited, and surprised. The six stickers are arranged in a relaxed slightly-overlapping random layout (not a rigid grid) across the {BACKGROUND_COLOR} background, each sticker rendered as a glossy puffy sticker with a clean thick white die-cut border around the character's silhouette and a subtle soft drop shadow underneath to suggest physical depth. The character must be identical across all six stickers — exact same line style, exact same color palette, exact same proportions, exact same outfit and hair from the uploaded reference; only the facial expression changes between each sticker. Add small motif elements (tiny hearts, stars, sparkles) scattered between the stickers as separate smaller puffy stickers in colors matched to the character's palette. Include one small text sticker rendered as a separate puffy sticker with the word {TEXT_WORD} in a rounded, pop, cute display font with a clean colorful gradient — placed naturally among the character stickers. The overall look is a printable Etsy / Redbubble sticker sheet — soft puffy glossy 3D feel, cute pop aesthetic, clean white die-cut borders. Single 4:5 portrait sticker-sheet image.

**Rules the AI must follow:**
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt
- Character consistency is the highest-priority constraint — match the uploaded reference's line style, color palette, proportions, outfit, hair, and accessories exactly across all six stickers; only the facial expression varies (the same character must be unmistakably recognizable in every sticker)
- All six expressions must be visibly distinct — neutral, smiling, angry, sad, happy / excited, surprised — no two stickers showing the same expression
- Render the exact text "{TEXT_WORD}" with exact spelling and exact letter count — no AI re-stylization of the word, no added punctuation, no added emoji, no shortened or extended variants
- Each sticker has a clean thick white die-cut border following the character / motif / text silhouette, with a subtle soft drop shadow — must read as a physical printable sticker, not a flat flat illustration
- Relaxed random layout — slight overlap between stickers is OK; do not arrange in a rigid grid, do not align to a uniform spacing
- Single image output — one 4:5 sticker sheet, not a contact sheet of 6 separate squares, not a variant grid
- No extra text, no captions, no watermarks, no logos beyond the one specified {TEXT_WORD} sticker
- Output the image directly without explaining the prompt back
- All text in English Latin script

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clean reference image of your character (PNG with transparent or simple background — this is the character the AI will keep identity-locked across all six expression variants)
- {TEXT_WORD} = SLAY (or pick one short word ≤8 letters — e.g., MOOD / CUTIE / VIBES / YASSS — see Bonus tips below)
- {BACKGROUND_COLOR} = soft warm cream (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single photoreal 4:5 portrait high-end editorial apparel lookbook mockup in clean minimal fashion-advertisement style. Background is a warm off-white to light khaki gradient studio backdrop, very minimal, soft, airy, with subtle warm tones. The main product is {GARMENT_TYPE_AND_COLOR}, ghost-hung on a silver metal clothing rack stand with a visible clamp at the top-right of the garment, front-facing, hanging naturally with slight realistic drape. On the center chest, the design from the uploaded reference image is rendered as if puff-embroidered or screen-printed onto the fabric — apply the uploaded design exactly as uploaded, preserving its colors, line weights, typography, and proportions, sized appropriately for a chest graphic (roughly 6–8 inches wide), with realistic fabric-interaction texture so the design follows the garment's micro-folds and is not floating like a sticker. Below the chest design, render the small text "{CHEST_TEXT}" in a clean sans-serif as a small puff-embroidery accent, exact spelling preserved. Add a small dark navy woven label patch on the right hem side seam and a visible dark navy neck tag at the collar. Typography overlay: top center of the frame shows the brand wordmark "{BRAND_NAME}" in a minimal clean sans-serif typeface, color-matched to the garment, kept small and tasteful (this is an editorial lookbook layout, not a billboard ad). Bottom one-fifth of the frame features three thin diagonal detail strips arranged horizontally: macro fabric texture of the garment material, macro closeup of the chest design embroidery/print stitching, and macro closeup of the dark navy neck label — all lit with the same warm studio lighting and integrated cleanly into the layout. Single 4:5 portrait editorial lookbook image, no person, no hands, no model.

**Rules the AI must follow:**
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt
- Design fidelity is the highest-priority constraint — match the uploaded design's colors, line weights, typography, and proportions exactly on the chest; do not redraw, restyle, recolor, or AI-interpret the design
- Render ALL on-image text exactly as specified — "{BRAND_NAME}" and "{CHEST_TEXT}" must appear with exact spelling, exact letter count, exact case, no swapped letters, no added swashes, no stylization beyond the specified typeface (apparel mockups fail most often at brand-text spelling — this is the #1 priority after design fidelity)
- Print / embroidery must interact realistically with the fabric — slight fabric-fold drape, embroidery has dimensional puff if specified, screen print sits flush with subtle texture absorption; no floating-sticker look
- Realistic fabric texture required — visible weave / knit / brushed fleece for hoodies; no over-polished waxy plastic surface
- Three diagonal detail strips at the bottom must be present, thin, and clearly readable — not optional, not replaced with a single big macro
- Single image output — no front/back grid, no contact sheet, no color-variant grid, no separate moodboard
- No extra brand logos, no captions, no watermarks, no price stickers (only the specified `{BRAND_NAME}` top-center and `{CHEST_TEXT}` on the chest)
- Output the image directly without explaining the prompt back
- All text in English Latin script

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: your chest design artwork (PNG of the print or embroidery graphic — the AI will apply it to the chest panel exactly as uploaded, preserving colors and detail)
- {GARMENT_TYPE_AND_COLOR} = an oversized vintage washed cream cotton-fleece pullover hoodie with a soft brushed interior and relaxed drop-shoulder fit (or pick one — see Bonus tips below for garment swaps)
- {BRAND_NAME} = NORTH STUDIOS (or your actual brand wordmark — short, ≤14 characters works best)
- {CHEST_TEXT} = est. 2024 (or any short tagline ≤12 characters — leave blank if you don't want chest text by replacing with the literal word "none" and the AI will omit it)

**Generate this image:**

A single photoreal 3:2 horizontal real-estate listing exterior photo of the front of the house shown in the uploaded reference image — preserve the home's actual silhouette, roofline, siding material, window count and placement, front door position, garage placement, porch or stoop, and overall yard layout exactly as uploaded; this must look like the same house, only cleaner and better-lit. Shoot from a standing eye-level position on the sidewalk or street directly in front of the property, framed at a slight 3/4 front angle (camera shifted roughly 15-20 degrees off dead-on so the home reads three-dimensional, not a flat elevation), with a 24mm-equivalent wide-angle lens — the real-estate industry standard that captures the home and front yard with spatial depth and no fish-eye distortion. The home occupies roughly 55-70% of the frame, anchored center, with breathing room of sky above and a clean strip of yard and walkway below. {TIME_OF_LIGHT} bathes the facade, casting soft natural shadows from architectural elements (eaves, window trim, porch overhang) and warm grazing highlights on the siding texture. The yard is clean and listing-ready — lawn freshly mowed with visible natural grass texture variation (real, not CGI carpet-perfect), walkway and driveway swept clear of debris, no parked cars in the driveway or street, no garbage cans visible, no hose or kids' toys, no political signs, no trash, no clutter of any kind. Landscaping is tidy but natural — trimmed shrubs and seasonal planting appropriate to the {STYLING_AESTHETIC}, real plant variation and natural irregularity (not symmetric plastic-perfect topiary). Sky is clear soft daylight with realistic clouds, white-balanced naturally and never over-saturated blue. Output a single 3:2 horizontal exterior listing-grade photo, photoreal, eye-level, true to the uploaded reference, curb-clean and re-lit.

**Rules the AI must follow:**
- Aspect ratio 3:2 horizontal — strict, locked at the start and the end of the prompt
- Architectural fidelity is the highest-priority constraint — match the uploaded photo's actual home silhouette, roofline, siding material, window count and placement, front door position, garage, porch, and yard layout exactly; do not invent a different house, do not change siding color, do not add or remove windows
- Realistic material texture required — visible real grass texture variation across the lawn, real concrete or paver micro-detail on the walkway and driveway, real brick or siding micro-imperfection on the facade, real wood grain on doors or trim if shown, real shingle or tile imperfection on the roof; no over-rendered CGI lawn carpet, no AI-perfect symmetric brick pattern, no glassy plastic siding
- No AI-perfect symmetry across the facade — preserve the home's actual real-world irregularity (one window slightly offset, eave shadow uneven, real-world architectural variation); mirrored CGI symmetry is the single biggest "fake-looking" tell
- No glossy CGI window reflections — windows show soft realistic daytime reflection of sky and nearby trees, not mirror-clean perfect glass; no fake fish-eye reflections inside the panes
- 24mm-equivalent wide angle only — no fish-eye distortion, no telephoto compression, no exaggerated low-angle hero perspective; the home must look spatially true-to-life, not stretched or warped
- Natural daylight only — no harsh blown-out sun, no impossibly saturated blue sky, no overdriven HDR halo around the roofline, no hot fake "magic-hour" glow; soft natural light, realistic clouds, true-to-life white balance
- Curb-clean — no parked cars, no garbage or recycling bins, no hose or kids' toys, no political signs, no trash, no laundry on a line, no construction debris, no for-sale sign in the yard, no people, no pets, no hands in frame
- No text, no captions, no watermarks, no on-image labels of any kind (no house number rendered as added overlay; if the real house has a visible number in the upload, preserve it as-is)
- Single image output — no contact sheet, no before/after split, no multiple angles in one frame
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears (e.g., a street sign in the background)

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear daytime photo of the front of your actual house from the sidewalk or street (this is what the AI will preserve — the home's real silhouette, roofline, siding, windows, front door, and yard layout; lighting and curb-cleanup happen on top)
- {TIME_OF_LIGHT} = soft late-afternoon golden-hour daylight from a slight front-side angle, warm gentle highlights on the facade (or pick one — see Bonus tips below for swaps)
- {STYLING_AESTHETIC} = suburban traditional with a clean front lawn, simple foundation shrubs, a tidy planted bed near the front door, and seasonal annuals in soft natural color (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single photoreal 3:2 horizontal real-estate listing photo of the living room shown in the uploaded reference image — preserve the room's actual architecture, window placement, ceiling height, fireplace or built-ins, and floor type exactly as uploaded; this must look like the same room, only cleaner and better-lit. Virtually stage the room with {FURNITURE_COMBINATION}, arranged in a comfortable, lived-in but uncluttered layout that respects the existing room geometry. Shoot at eye level with a 24mm-equivalent wide angle (the real-estate industry standard — spacious feeling without fish-eye distortion), camera anchored so the room's longest sightline runs into the frame. {TIME_OF_DAY} pours through the existing windows, casting soft natural shadows and warm highlights on the floor; interior lights are off so the daylight does the work. Modern, warm, elegant residential mood; earth tones; realistic materials; visible everyday-life details like a folded throw blanket and one or two styled coffee-table books. Negative space and breathing room around the seating area — no clutter, no personal items, no laundry, no cables. Output a single 3:2 horizontal listing-grade photo of the room, photoreal, eye-level, true to the uploaded reference.

**Rules the AI must follow:**
- Aspect ratio 3:2 horizontal — strict, locked at the start and the end of the prompt
- Architectural fidelity is the highest-priority constraint — match the uploaded photo's actual room shape, window count and position, ceiling height, fireplace / built-ins, and floor type exactly; do not invent a different room
- Realistic material texture required — visible wood grain on flooring, real fabric folds on sofa cushions, real subtle wall-paint micro-imperfections; no over-rendered CGI surfaces, no AI-perfect symmetry, no glossy waxy "showroom" highlights
- Natural daylight only — no harsh direct sun, no hot blown-out windows, no over-saturated colors; soft shadows, warm highlights, true-to-life white balance
- No cluttered interior, no personal items visible (no laundry, no cables, no half-finished coffee mugs, no toys, no family photos with faces) — clean styled-for-listing presentation
- No people, no hands, no pets in the frame
- No text, no captions, no watermarks, no on-image labels of any kind
- Single image output — no contact sheet, no before/after split, no multiple angles in one frame
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears (e.g., book spines on a shelf)

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear iPhone photo of your actual living room (this is what the AI will preserve — the room's real architecture, windows, fireplace, and floor; furniture will be virtually staged on top)
- {FURNITURE_COMBINATION} = a curved low-back sofa in soft oatmeal linen, a round walnut coffee table, a textured wool rug in cream, a single tall fiddle-leaf fig in a ceramic planter, a folded throw blanket in warm camel (or pick one — see Bonus tips below for swaps)
- {TIME_OF_DAY} = soft late-morning daylight (or pick one: bright midday daylight / warm golden-hour afternoon light / soft overcast diffused daylight)

**Generate this image:**

A single photoreal 3:2 horizontal real-estate listing photo of the kitchen shown in the uploaded reference image — preserve the kitchen's actual cabinet layout, island position and shape, sink location, appliance placement (range / oven / fridge / dishwasher), window placement, ceiling height, and floor type exactly as uploaded; this must look like the same kitchen, only cleaner and better-styled. Apply a {STYLE_AESTHETIC} aesthetic: countertops are wiped clean and styled with only one of the following — a single styled wooden bowl of fresh whole fruit, OR a clean wooden cutting board with a single trimmed potted herb, OR a single matte-finish coffee maker — and absolutely nothing else on the counters. Sink is empty and dry; no dishes, no sponges, no soap bottles visible. Island holds at most one small styled element (a small folded linen tea towel or a single cookbook propped open). No mail, no keys, no chargers, no fridge magnets, no family photos, no small appliances cluttering every counter, no exposed cleaning products, no dish towels draped on handles. Shoot at eye level with a 24mm-equivalent wide angle (the real-estate industry standard — spacious feel without fish-eye distortion), camera anchored so the kitchen island (or longest counter run if no island) is the visual anchor with the window in or near the frame. {TIME_OF_DAY} pours through the existing windows, casting soft natural shadows and warm highlights on the counters and floor; interior lights and pendant lights stay off so the daylight does the work. Modern, warm, inviting residential mood; earth tones grounded by whatever cabinet color and floor type are shown in the reference; realistic materials; visible everyday-life details kept to the single styled object only. Output a single 3:2 horizontal listing-grade kitchen photo, photoreal, eye-level, true to the uploaded reference.

**Rules the AI must follow:**
- Aspect ratio 3:2 horizontal — strict, locked at the start and the end of the prompt
- Architectural fidelity is the highest-priority constraint — match the uploaded photo's actual cabinet layout, island position and shape, sink location, appliance placement, window count and position, ceiling height, and floor type exactly; do not invent a different kitchen, do not move the island, do not change cabinet color unless explicitly requested in the style aesthetic
- Realistic material texture required — real wood grain on cabinets or flooring if shown, real subtle micro-scratches and soft fingerprint texture on stainless steel appliances, real grout lines on tile, real subtle wall-paint micro-imperfections; no over-rendered CGI stainless mirror reflections, no AI-perfect symmetry across the island, no glassy waxy "showroom" countertop highlights
- Natural daylight only — no harsh direct sun, no hot blown-out windows, no over-saturated colors; soft shadows, warm highlights, true-to-life white balance
- Listing-clean counters and sink — no dishes in sink, no dish towels draped on oven or cabinet handles, no exposed cleaning products, no small appliance clutter beyond the single styled object, no fridge magnets, no mail, no keys, no chargers, no personal photos
- No people, no hands, no pets in the frame
- No text, no captions, no watermarks, no on-image labels of any kind (no brand logos on appliances if avoidable)
- Single image output — no contact sheet, no before/after split, no multiple angles in one frame
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears (e.g., cookbook spine on island)

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear iPhone photo of your actual kitchen (this is what the AI will preserve — the real cabinet layout, island position, appliances, windows, and floor; styling and lighting will be cleaned up on top)
- {STYLE_AESTHETIC} = white-bright modern with crisp white shaker cabinets, light quartz countertops, brushed nickel hardware, and whatever flooring is shown in the reference (or pick one — see Bonus tips below for swaps)
- {TIME_OF_DAY} = soft late-morning daylight (or pick one: bright midday daylight / warm golden-hour afternoon light / soft overcast diffused daylight)

**Generate this image:**

A single photoreal 3:2 horizontal real-estate listing photo of the {BEDROOM_TYPE} shown in the uploaded reference image — preserve the room's actual architecture, window placement, bed position and orientation, ceiling height, closet doors, and floor type exactly as uploaded; this must look like the same room, only cleaner and better-styled. Apply a {STYLE_AESTHETIC} aesthetic: a freshly made bed with crisp wrinkle-free linen featuring one or two intentional natural soft folds (lived-in but not messy), two styled pillows plus one accent cushion, a single small nightstand lamp on, one nightstand holds only a small styled book or a single trimmed potted plant (no chargers, no phones, no personal photos, no medication, no clutter of any kind). Shoot at eye level with a 24mm-equivalent wide angle (real-estate industry standard — spacious feel without fish-eye distortion), camera anchored so the bed is the visual anchor with the window in or near the frame. {TIME_OF_DAY} fills the room through open curtains or blinds, casting gentle natural shadows and warm highlights on the floor and bed linen; interior lights are off so the natural light does the work. Modern, warm, restful residential mood; earth tones; realistic materials; warm oak or whatever flooring is shown in the reference, off-white or whatever wall color is shown in the reference. Negative space and breathing room — no clutter, no personal items, no laundry on the chair, no cables visible. Output a single 3:2 horizontal listing-grade bedroom photo, photoreal, eye-level, true to the uploaded reference.

**Rules the AI must follow:**
- Aspect ratio 3:2 horizontal — strict, locked at the start and the end of the prompt
- Architectural fidelity is the highest-priority constraint — match the uploaded photo's actual room shape, window count and position, bed location and orientation, ceiling height, closet doors, and floor type exactly; do not invent a different bedroom or move the bed
- Realistic material texture required — real fabric wrinkle texture on linen, real wood grain on flooring, real plant leaf variation if any plant is shown, real subtle wall-paint micro-imperfections; no over-rendered CGI smoothness, no AI-perfect symmetry, no glossy waxy bed-linen highlights
- Natural daylight only — no harsh direct sun, no hot blown-out windows, no over-saturated colors; soft shadows, warm highlights, true-to-life white balance
- Listing-clean nightstand and surfaces — no personal items anywhere (no chargers, no phones, no medication, no half-full water glasses, no family photos with faces, no laundry on chairs)
- No people, no hands, no pets in the frame
- No text, no captions, no watermarks, no on-image labels of any kind
- Single image output — no contact sheet, no before/after split, no multiple angles in one frame
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears (e.g., book spine on nightstand)

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear iPhone photo of your actual bedroom (this is what the AI will preserve — the room's real architecture, bed position, windows, and floor; styling and linen will be cleaned up on top)
- {BEDROOM_TYPE} = master bedroom (or pick one: guest bedroom / kids bedroom / second bedroom / primary suite)
- {STYLE_AESTHETIC} = modern minimalist with warm oak wood flooring and off-white walls, calm restful neutral palette (or pick one — see Bonus tips below for swaps)
- {TIME_OF_DAY} = soft late-morning daylight (or pick one: bright midday daylight / warm golden-hour afternoon light / soft overcast diffused daylight)

**Generate this image:**

A single photoreal 4:5 portrait studio product shot of a single {BAG_TYPE} standing upright, centered on a clean white seamless background with soft even daylight from the upper-left and a subtle realistic floor shadow. The bag is shown full and gently bulging as if filled with merchandise, with the bag opening in a standard retail format at the top. On the front center of the bag, render the exact logo from the uploaded reference image — match its shape, proportions, wordmark, typography, and any tagline exactly as uploaded, sized to occupy roughly 35-45% of the bag's front face. The logo appears as if screen-printed flat onto the bag surface, with crisp edges, high contrast against the bag color, and slight conformity to the bag's surface curvature, soft folds, and crinkles. The bag body shows realistic crinkles, soft folds, gentle highlights, and natural material texture from the contents inside — not a flat digital mockup. Photographed at eye level so the bag looks true-to-life. No other objects, no hands, no clutter, no background props. Single 4:5 portrait studio shot, product-only, eye-level, hero-centered.

**Rules the AI must follow:**
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt
- Logo fidelity is the highest-priority constraint — match the uploaded logo's shape, proportions, wordmark, typography, and tagline exactly; do not redraw the brand name in a different font, do not invent additional graphic elements, do not change the logo's color
- Realistic material texture required — visible crinkles, soft folds, gentle highlights, natural bag-stretch from the contents inside; no over-polished flat digital mockup look, no waxy "AI plastic" surface
- The screen-printed logo must subtly conform to the bag's surface — slight bending across folds, very subtle ink-on-fabric or ink-on-plastic texture; not floating perfectly flat above the surface
- Single image output — no contact sheet, no variant grid, no before/after split, no multiple angles in one frame
- No additional text, captions, watermarks, price tags, or background decoration beyond what already exists in the uploaded logo
- Output the image directly without explaining the prompt back
- All text in English Latin script

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clean version of your logo (SVG, PNG with transparent background, or high-resolution JPG — this is the exact mark the AI will render onto the bag)
- {BAG_TYPE} = glossy black plastic merchandise bag with a flat-folded top opening (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single photoreal 1:1 square close-up macro shot of the exact logo from the uploaded reference image rendered as a {PROCESS_TYPE} relief. Match the uploaded logo's shape, proportions, wordmark, and typography exactly — the relief should be your logo, not a substituted generic mark. Show the logo pressed into or out of the substrate with visible three-dimensional unevenness, soft shadow falling into the depressed areas, and gentle highlights catching the raised edges. The substrate fills the frame as the full background — no rounded card frame, no business-card border, no generic luxury-template overlay. Lighting is hard raking side-light from the lower-left at a 15-degree grazing angle, which is the lighting that makes the relief readable — strong highlight on one edge of each pressed shape, soft shadow on the opposite edge, with the rest of the surface in soft mid-tone so the texture stays the hero. Slight natural imperfection in the substrate is welcome (paper fiber, micro-grain in the metal, gentle stock texture) — this is what makes it look like a real photographed sample instead of a flat vector mockup. Single 1:1 square macro relief shot, substrate-only, top-down or near-top-down angle.

**Rules the AI must follow:**
- Aspect ratio 1:1 square — strict, locked at the start and the end of the prompt
- Logo fidelity is the highest-priority constraint — match the uploaded logo's shape, proportions, wordmark, and typography exactly; do not redraw, re-letter, or substitute the mark
- Hard raking side-light required — directional light from the lower-left at 15-degree grazing angle, with clear highlight-on-one-edge / shadow-on-opposite-edge across every pressed shape; no flat front-lighting (flat light makes embossing invisible)
- Show the relief on the substrate as the unframed mark — no rounded business-card frame, no generic "luxury template" border, no stock-photo card-shape overlay around the logo
- Realistic substrate texture required — visible paper fiber / metal micro-grain / wax flow / letterpress bite; no over-polished flat digital look
- Single image output — no contact sheet, no variant grid, no before/after split
- No additional text, captions, or watermarks beyond what already exists in the uploaded logo
- Output the image directly without explaining the prompt back
- All text in English Latin script

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clean version of your logo (SVG, PNG with transparent background, or high-resolution JPG — this is the exact mark the AI will render in relief)
- {PROCESS_TYPE} = embossed (raised) into thick cream cotton paper stock with a soft natural fiber texture (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single photoreal 4:5 portrait ultra-luxury campaign hero shot featuring the exact box from the uploaded reference image as the centered hero — match its silhouette, proportions, color, label, finish, and brand wordmark exactly. The box, with a {BOX_FINISH} finish, sits elevated on a sculptural pedestal at the center of an architectural vault-like chamber, as if it belongs to a private collection unavailable to the public. Surrounding the hero box are layered compartments, recessed niches, sliding panels, and polished display surfaces forming a custom-built chamber around it. Three to five supporting artifacts are arranged in the surrounding niches — clearly secondary in size and lighting, never competing with the hero box for attention. The chamber materials are velvet-lined cavities, brushed brass inlays, smoked glass panels, and lacquered wood — locked, do not substitute marble / concrete / plastic. The overall mood is {CHAMBER_MOOD}, with pinpoint specular highlights on the precious materials, dramatic edge light grazing across the hero box, soft shadow falloff into the surrounding niches, and velvet-deep blacks. Single 4:5 portrait luxury campaign hero shot, hero-box anchored center-frame on its pedestal.

**Rules the AI must follow:**
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt
- Box fidelity is the highest-priority constraint — match the uploaded box's silhouette, proportions, color, label artwork, finish, and brand wordmark exactly; do not invent a different box or redraw the label
- Maximum 3 supporting artifacts in the surrounding compartments — each clearly secondary in size and lighting; never crowd the niche with random objects, books, sculptures, or duplicate bottles
- Lock the four chamber materials — velvet + brushed brass + smoked glass + lacquered wood — do not substitute marble, concrete, plain glass, or stainless steel
- Realistic material texture required — visible velvet pile, brass brushed grain, glass refraction, wood grain in the lacquer; no over-polished AI-plastic surface, no waxy highlights
- Single image output — no contact sheet, no variant grid, no before/after split
- No additional text, captions, watermarks, or signage in the chamber beyond what already exists on the uploaded box
- Output the image directly without explaining the prompt back
- All text in English Latin script

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clean image of your box (a flat dieline, 3D render, prototype photo, or label artwork mocked onto a box silhouette — the AI matches the uploaded reference exactly)
- {BOX_FINISH} = matte black premium with subtle metallic accents (or pick one — see Bonus tips below for swaps)
- {CHAMBER_MOOD} = high-jewelry campaign meets private museum exhibition, cool refined tones (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single photoreal 4:5 portrait studio product shot of a single stand-up zipper-top pouch standing upright on a clean neutral surface with soft natural daylight from the upper-left and a subtle realistic floor shadow. The pouch is a {POUCH_FINISH} stand-up flexible pouch with a clearly visible zipper top (resealable plastic zipper with a small zipper pull tab), a flat-bottom gusset at the base that lets it stand upright on its own, visible side seals running vertically along both edges, and a small triangular tear notch at the top right corner just below the zipper. The pouch is shown full and gently rounded from the contents inside — not a flat empty bag, not over-stuffed. On the front face of the pouch, render the exact logo from the uploaded reference image — match its shape, proportions, wordmark, typography, and any tagline exactly as uploaded, positioned per the {LABEL_LAYOUT}, sized appropriately for the pouch front. The printed surface of the pouch shows realistic matte ink-on-mylar print quality — the logo edges are crisp but conform subtly to the soft folds and gentle curvature of the pouch body, with the printed surface having the same matte texture as the surrounding pouch material (no separate floating sticker look). The pouch body shows realistic soft folds, gentle highlights where the surface catches light, slight natural irregularity in how the front face curves, and the visible micro-texture of the matte mylar substrate — not a flat digital mockup, not a perfect cylinder. Photographed at eye level so the pouch looks true-to-life. No other objects, no hands, no clutter, no background props beyond the clean neutral surface and its soft floor shadow. Single 4:5 portrait studio shot, product-only, eye-level, pouch hero-centered.

**Rules the AI must follow:**
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt
- Logo fidelity is the highest-priority constraint — match the uploaded logo's shape, proportions, wordmark, typography, and any tagline exactly; do not redraw the brand name in a different font, do not invent additional graphic elements, do not change the logo's color, do not add nutrition claims or marketing copy beyond what is in the uploaded logo
- Render the pouch anatomy exactly — visible plastic zipper across the top with a small pull tab, a flat-bottom gusset at the base that lets the pouch stand upright on its own, visible vertical side seals along both edges, and a small triangular tear notch at the upper right just below the zipper; do not render a generic soft bag with no zipper, no gusset, or no seals
- Realistic matte mylar texture required — visible soft folds, gentle highlights, slight natural curvature on the front face, micro-texture of the substrate; no over-polished AI-perfect glossy plastic, no mirror highlights, no cartoon shine, no waxy showroom finish
- The printed logo must subtly conform to the pouch's surface — slight bending across folds and curvature, ink-on-mylar texture matching the surrounding matte substrate, not floating perfectly flat above the surface as a separate layer
- Single image output — no contact sheet, no variant grid, no before/after split, no multiple angles in one frame
- No additional text, captions, watermarks, nutrition labels, ingredient lists, weight callouts, or background decoration beyond what already exists in the uploaded logo
- Output the image directly without explaining the prompt back
- All text in English Latin script

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clean version of your logo (SVG, PNG with transparent background, or high-resolution JPG — this is the exact mark the AI will render onto the pouch front face)
- {POUCH_FINISH} = matte kraft-brown stand-up pouch with a soft natural fiber-textured surface (or pick one — see Bonus tips below for swaps)
- {LABEL_LAYOUT} = centered on the upper-middle third of the pouch front, occupying roughly 50-60% of the front face width, with clean breathing room above and below (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single 9:16 vertical phone wallpaper with a calm, minimal aesthetic, designed to sit cleanly behind iOS or Android lock-screen UI. The image features {SUBJECT} as the main element, sitting in the middle band of the frame, with generous negative space around it. A subtle smooth {GRADIENT_PALETTE} top-to-bottom gradient washes across the full canvas as the base, giving the wallpaper a unified atmospheric backdrop without any visible color banding. The overall feel is {MOOD}, quiet and restful — the kind of wallpaper that doesn't fight for attention when you glance at your phone. The top one-third of the frame is reserved as calm negative space for the iOS clock and date area, and the bottom one-fifth is also kept calm and uncluttered for the icon dock — no important detail of the subject should fall inside those zones. No people, no text, no logos, no decorative borders. Single 9:16 vertical wallpaper, calm and minimal.

**Rules the AI must follow:**
- Aspect ratio 9:16 vertical — strict, locked at the start and the end of the prompt
- The top one-third and the bottom one-fifth of the frame must stay as calm negative space — no important detail, no busy texture, no high-contrast element in those zones (these areas sit behind the clock and the icon dock)
- No blur effects, no atmospheric haze, no soft-focus across the whole image — the subject and gradient should both be cleanly rendered
- No cropping of the subject — the main element sits comfortably inside the frame with breathing room on all sides
- No text, no captions, no watermarks, no logos, no UI mockups anywhere in the image
- Realistic detail in the subject — natural unevenness, micro-detail, no flat plastic "AI render" look
- Single image output — no contact sheet, no variant grid, no before/after split
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage somehow appears

**Replace these placeholders with your details:**
- {SUBJECT} = a single minimalist mountain silhouette at dawn (or pick one — see Bonus tips below for swaps)
- {GRADIENT_PALETTE} = soft peach-pink fading into pale lavender (or pick one — see Bonus tips below for swaps)
- {MOOD} = calm, soft, slightly dreamy

**Generate this image:**

A single 9:16 vertical photoreal wildlife wallpaper of {ANIMAL}, viewed head-on with an intense, curious gaze. The animal is peeking over a textured natural ledge in {BACKGROUND_TONE}, with its front paws resting symmetrically on either side of its face. Its large eyes — sharp, alert, full of wonder — are positioned in the lower two-thirds of the frame so that the top one-third stays clean negative space for the iOS clock and date. Rounded ears are perked up, catching the light. Every fur strand is individually visible, with realistic micro-texture across the face and paws, fine whiskers catching the {TIME_OF_DAY} light. Soft, even {TIME_OF_DAY} lighting falls naturally across the animal, with critically sharp focus on the eyes and face. Shallow depth of field renders the background as a smooth, uniform bokeh in the same {BACKGROUND_TONE} tone, free of any distracting detail, so the animal stays the unambiguous hero of the frame. Professional wildlife portrait quality, natural yet intimate feel. Single 9:16 vertical photo, photoreal.

**Rules the AI must follow:**
- Aspect ratio 9:16 vertical — strict, locked at the start and the end of the prompt
- The top one-third of the frame must stay clean negative space — no fur, no paws, no detail there (this sits behind the iOS clock and date)
- Realistic fur texture is mandatory — individually visible fur strands, fine whiskers, natural micro-imperfection in the eye reflections and nose moisture; absolutely no smoothed-over "AI plastic" / "rendered CGI" fur, no glossy waxy highlight, no over-perfect symmetry
- Critically sharp focus on the eyes and face — eye catchlight visible, slight asymmetry in the pupils OK
- Background must be a smooth uniform bokeh — no leaves, no branches, no patterns, no rocks visible in the blurred zone (anything visible in the background fights the home-screen icons)
- No people, no text, no captions, no watermarks, no logos anywhere in the frame
- Single image output — no contact sheet, no variant grid, no before/after comparison
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears

**Replace these placeholders with your details:**
- {ANIMAL} = a baby tiger cub with striking orange-gold fur and dark black stripes (or pick one — see Bonus tips below for swaps)
- {BACKGROUND_TONE} = deep forest green
- {TIME_OF_DAY} = soft morning daylight

**Generate this image:**

A single 9:16 vertical dreamcore-surrealism phone wallpaper. The main scene is {SCENE}, occupying the lower two-thirds of the frame with clean sweeping curves and ultra-sharp natural texture detail (visible sand grain, ice crystal, or wave detail depending on the scene). Above the scene, the upper third of the sky is kept as calm negative space, dotted only with a handful of minimalist tiny stars — sparse, not dense — so that the iOS clock and date can sit cleanly on top without competing with sky detail. The color palette leads with a {PALETTE_PRIMARY}, supported by midnight navy blue in the sky and a soft gold rim of light tracing the edges of the main landscape. The lighting source — a large soft moon, an aurora ribbon, or whatever celestial element the scene calls for — sits in the upper portion of the frame, casting gentle rim-light along the curves of the landscape below. The overall feel is {MOOD}, photoreal but dreamlike, cinematic 8K UHD clarity. Single 9:16 vertical wallpaper, photoreal and surreal.

**Rules the AI must follow:**
- Aspect ratio 9:16 vertical — strict, locked at the start and the end of the prompt
- The upper third of the frame must stay sparse — only a small number of tiny minimalist stars (or one celestial element like the moon / aurora) — no dense star fields, no nebulae, no clouds (this is where the iOS clock and date sit)
- The main landscape sits in the lower two-thirds of the frame with sweeping clean curves — no clutter, no scattered rocks, no foreground debris
- Ultra-sharp natural texture detail in the landscape — visible sand grain / ice crystal / water ripple appropriate to the scene; no smoothed-over flat AI surface
- High depth of field — both the landscape and the sky in clear focus, not a shallow-DoF portrait blur
- No people, no animals, no text, no captions, no watermarks, no logos anywhere in the frame
- Single image output — no contact sheet, no variant grid, no before/after split
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage somehow appears

**Replace these placeholders with your details:**
- {SCENE} = smooth flowing sand dunes under a massive glowing crescent moon (or pick one — see Bonus tips below for swaps)
- {PALETTE_PRIMARY} = deep terracotta orange
- {MOOD} = dreamy, surreal, quietly cosmic

**Generate this image:**

A single 9:16 vertical lofi flat-vector phone wallpaper. The composition is a clean three-zone vertical stack: the top half of the frame is pure white empty negative space (reserved for the iOS clock and date), the middle band features {SUBJECT_TYPE} from the uploaded reference image — drawn as a minimal flat-vector line-art illustration lying on its side, resting its head on the top edge of a large rounded rectangle banner, with closed eyes, smooth curves, and a relaxed sleepy pose — and the bottom one-third of the frame is occupied by a soft pastel {BANNER_COLOR} rounded rectangle banner with a subtle soft shadow underneath, sized to sit cleanly behind the iOS home-screen icon dock. The line art preserves the recognizable silhouette and proportions of the subject from the uploaded reference (so it still looks like the same animal / plushie / character). Lines are thin, dark gray to soft black, crisp vector, no shading inside the subject. The overall mood is playful, calm, and quiet — modern UI illustration aesthetic, the kind of wallpaper that makes you smile at the lock screen. Single 9:16 vertical wallpaper, flat vector.

**Rules the AI must follow:**
- Aspect ratio 9:16 vertical — strict, locked at the start and the end of the prompt
- Three-zone vertical composition is mandatory — top half pure white empty (clock zone), middle band sleeping subject on banner edge, bottom one-third pastel banner (icon dock zone); no decorative elements drift into the top half
- The subject must be recognizable as the subject from the uploaded reference image — preserve its silhouette, proportions, and distinguishing features (ear shape / tail / pattern / hairstyle) in the line-art simplification
- Closed eyes, smooth curves, relaxed sleepy pose are all required — no open eyes, no standing pose, no action pose
- No text, no captions, no quotes, no typography anywhere on the banner or in the white space — buyers will add their own widget or quote on top in the iOS lock-screen edit UI
- Crisp vector line art — clean thin outlines, no painterly brush, no gradient shading inside the subject; the banner can have a single subtle soft shadow only
- Single image output — no contact sheet, no variant grid, no before/after split
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage somehow appears

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear photo of your pet, plushie, or character (this is the subject the AI will turn into the sleeping line-art figure — without it the result will be a generic stock cartoon)
- {SUBJECT_TYPE} = a cat (or pick one — see Bonus tips below: dog / plushie bear / anime OC / rabbit / etc.)
- {BANNER_COLOR} = soft pastel lemon green

**Generate this image:**

A single 3:4 vertical vintage patent document page in the style of a late 1800s United States Patent Office filing, featuring the exact object from the uploaded reference image rendered as Victorian-era fountain-pen technical illustration. The page presents three labeled views of the uploaded object: Fig. 1 (front elevation), Fig. 2 (side profile), and Fig. 3 (exploded or cutaway view showing internal mechanism) — each numbered with small leader lines pointing to key components. Preserve the object's distinctive silhouette, proportions, and component breakdown exactly as in the uploaded reference photo — the object must remain instantly recognizable. The drawings are rendered in fountain-pen black ink line-art with cross-hatch shading on aged ivory paper (3:4 vertical), {AGING_INTENSITY} foxing stains, soft horizontal fold creases, slightly torn lower edge. Handwritten cursive annotations in faded brown fountain-pen ink describe each component (simulated text — illegible scribble is fine), with a hand-lettered title "{OBJECT_LABEL}" near the top in {ERA_VARIANT} display lettering. An embossed circular seal and red wax stamp appear in one corner; a signed inventor's name and date in flowing cursive appear at the bottom. The entire page feels like a recovered archival document — authoritative, historic, slightly mysterious, with real fountain-pen ink bleed, faint line wobble, and hand-pressed pressure variation. Single 3:4 vertical patent document image, object-fidelity preserved.

**Rules the AI must follow:**
- Aspect ratio 3:4 vertical — strict, locked at the start and the end of the prompt
- Object fidelity is the highest-priority constraint — the silhouette, proportions, and identifiable parts of the uploaded reference object must remain recognizable; do not substitute a generic Victorian invention
- Render "Fig. 1", "Fig. 2", "Fig. 3" and the title "{OBJECT_LABEL}" as legible English Latin script — every other text element (annotations, signature, date, seal lettering) is simulated handwritten scribble allowed to be illegible
- Hand-drawn Victorian look required — real fountain-pen ink bleed, slight line wobble, hand-pressed pressure variation; no perfect CAD vector lines, no modern industrial-design aesthetic, no clean digital line-art
- Aged paper realism required — visible foxing stains, fiber texture, fold creases, slightly uneven edges; no flat AI plastic paper, no over-clean white background
- Single image output — no multi-page document, no contact sheet, no before/after split, no grid of variants
- Output the image directly without explaining the prompt back
- All text in English Latin script — no substitution to other writing systems

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit photo of the object you want rendered as a patent (without this the AI has nothing to stylize — it will invent a generic Victorian gadget instead of YOUR thing)
- {OBJECT_LABEL} = Improved Mechanical Coffee Brewer (or whatever you want the patent title to say — keep it short, 3-6 words, Victorian-sounding)
- {ERA_VARIANT} = late 1800s US Patent Office (or pick one — see Bonus tips below for swaps)
- {AGING_INTENSITY} = moderate (or "light" for a cleaner crisp look, "heavy" for a beat-up archival feel)

**Generate this image:**

A single 3:4 vertical vintage encyclopedia field-guide poster in the style of a 1800s Victorian natural-history atlas, with the subject from the uploaded reference image rendered as the central engraved illustration. Preserve the uploaded subject's species/breed-identifying features, distinctive markings, body proportions, and overall pose exactly as in the reference photo — the subject must remain instantly recognizable. The central illustration occupies roughly 45% of the page (slightly upper-center), rendered as a detailed Victorian engraved drawing with fine cross-hatch shading, ink stippling, and hand-drawn line quality. An engraved-style header at the top reads "{SUBJECT_NAME}" in Victorian display lettering. Surrounding the central illustration are exactly 10 labeled infographic panels arranged in a 2-column grid (5 panels on the left of the central illustration, 5 on the right), each with a small Victorian-engraved icon and a legible English header — the 10 default labels are: Species Type, Anatomy, Habitat Region, Diet, Lifecycle, Temperament, Distinctive Markings, Notable Behaviors, Field Notes, Specimen Date. Body text inside each panel is simulated handwritten scribble (illegible is fine — only the headers must be readable). The background is aged parchment paper with a {COLOR_PALETTE} color palette, soft burnt edges, foxing stains, and faint old-map texture. The poster is in {POSTER_THEME} style, with engraved ornamental borders, small ancient symbols in corners, and a Victorian cartouche frame around the central illustration. Single 3:4 vertical poster, subject-fidelity preserved, 10 panels hard-locked.

**Rules the AI must follow:**
- Aspect ratio 3:4 vertical — strict, locked at the start and the end of the prompt
- Subject fidelity is the highest-priority constraint — the uploaded subject's recognizable features (face / breed / markings / proportions) must remain identifiable; do not substitute a generic creature
- Exactly 10 infographic panels — no more, no fewer; the 10 panel headers and "{SUBJECT_NAME}" must render as legible English Latin script; all body text inside panels is simulated scribble allowed to be illegible
- Hand-engraved Victorian look required — visible cross-hatch shading, ink stippling, slight line wobble, real engraved-plate imperfections; no clean digital vector lines, no flat minimalism, no modern UI styling
- Aged paper realism required — foxing stains, fiber texture, faint fold creases, burnt edges; no flat AI plastic background
- Single image output — no contact sheet, no multi-page document, no variant grid
- Output the image directly without explaining the prompt back
- All text in English Latin script — no substitution to other writing systems

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear photo of the subject you want as the field-guide entry (without this the AI has nothing to stylize — it will invent a generic Victorian creature instead of YOUR subject)
- {SUBJECT_NAME} = Felis Domesticus — the Common House Cat (or whatever fits your subject — Latin-sounding scientific name + common name works best)
- {POSTER_THEME} = Victorian natural-history atlas (or pick one — see Bonus tips below for swaps)
- {COLOR_PALETTE} = sepia and faded gold (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single 3:4 vertical Japanese Edo-period Ukiyo-e woodblock print reimagining the scene from the uploaded reference photo as a surreal collaboration between Hokusai and Hiroshige. Preserve the uploaded photo's composition exactly — the number of people, their relative positions, what they are doing, their clothing/identifying objects, and the key recognizable setting elements (buildings / landmarks / props) must all match the reference. The scene shown is {SCENE_DESCRIPTION}. Apply Edo transformation logic to every modern element: smartphones become glowing illustrated paper scrolls; cars and bikes become ox-drawn wooden carriages or painted-paper palanquins; metro trains and buses become giant articulated wooden centipede-carriages; skyscrapers and apartment buildings become endless towering wooden pagodas reaching into dramatic clouds; modern clothing becomes Edo-era kimono and yukata in the same colors/patterns; cameras and screens become painted bamboo lens-boxes; cafes and shops become wooden lantern-lit market stalls. The composition uses flattened Ukiyo-e perspective with large, bold, hand-carved ink outlines — no Western chiaroscuro, no realistic shadows. The background features heavily stylized wave patterns and dramatic swirling clouds with a distant Mt. Fuji on the horizon. The {COLOR_INTENSITY} palette is strictly limited to traditional Edo mineral pigments — Prussian blue dominant, vermilion red, muted yellow ochre, soft black ink, off-white washi paper. Texture: strong visible wood grain across the entire surface, rough washi paper fibers, pigment bleeding at color boundaries, slight 1-2mm color-plate misalignment between blue and red plates as if hand-pressed. Lighting is soft, flat, shadow-free with no digital gradients. Include vertical Japanese-style calligraphy text down one side describing the scene (simulated brushstroke text is fine) and a traditional red square artist seal stamp in the lower corner reading "{SEAL_TEXT}". Single 3:4 vertical Ukiyo-e woodblock print.

**Rules the AI must follow:**
- Aspect ratio 3:4 vertical — strict, locked at the start and the end of the prompt
- Composition fidelity is the highest-priority constraint — preserve the number of people, their positions, their actions, their clothing/identifying objects, and recognizable setting landmarks from the uploaded reference photo; faces will be Ukiyo-e-stylized (accepted aesthetic trade-off) but everything else must anchor to the upload
- "{SEAL_TEXT}" on the red artist seal must render as legible English Latin script — all other text (vertical calligraphy down the side) is simulated brushstroke marks allowed to be illegible
- Physical woodblock print look required — visible wood grain texture, real pigment bleeding at color boundaries, slight color-plate misalignment, rough washi paper fiber; no clean digital gradients, no smooth shading, no Photoshop AI plastic look
- Strict Edo color palette — Prussian blue / vermilion red / muted yellow ochre / soft black / off-white washi only; no neon, no Western pastels, no full-spectrum digital colors
- Flattened Ukiyo-e perspective only — no Western 3D depth, no realistic cast shadows, no photographic lighting
- Single image output — no multi-panel comic layout, no contact sheet, no variant grid
- All meaningful text in English Latin script — applies to `{SEAL_TEXT}` on the artist seal; decorative vertical brushstroke calligraphy down the side is allowed to be stylized illegible marks (intentional Ukiyo-e exception to §4 #6)
- Output the image directly without explaining the prompt back

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear photo of the scene you want re-imagined (without this the AI has nothing to anchor to — it will invent a generic Edo scene instead of YOUR moment)
- {SCENE_DESCRIPTION} = our family of four crossing a busy city street at dusk (one short sentence describing what's in your uploaded photo — helps the AI interpret the image correctly)
- {SEAL_TEXT} = SC (2-4 letters or short initials for your artist seal — keep it short, this is a square stamp)
- {COLOR_INTENSITY} = traditional muted Edo mineral pigments (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single 3:4 portrait professional high-resolution paint-by-numbers template generated from the uploaded reference photo, suitable for A4 portrait printing at 300 DPI. The main artwork is a clean black-and-white line drawing on a pure white background — only black outlines, black numbers inside regions, and white unfilled paintable areas. Analyze the uploaded image and convert it into simplified, fully closed, easy-to-paint regions while preserving the original composition, main objects, faces, recognizable text/logos, foreground, background, and key details — the original subject must remain instantly recognizable as "{SUBJECT_NAME}". Use {COLOR_COUNT} colors total and a {COMPLEXITY_LEVEL} region count. Every paintable white region contains the correct black number, centered clearly inside the matching area, upright, readable at print resolution, and not touching the outlines. Region outlines are clean continuous black strokes, all regions fully closed (no broken segments). At the bottom of the same image, add a neat legend section with the heading "Instructions", the sentence "Paint the areas with the matching colors.", and a numbered list of every color used with a small filled color swatch next to each number and the English color name. The page title "{SUBJECT_NAME}" appears as small clean text at the top center. Single 3:4 portrait paint-by-numbers page, line-art main artwork plus bottom legend, print-ready.

**Rules the AI must follow:**
- Aspect ratio 3:4 portrait (A4-compatible) — strict, locked at the start and the end of the prompt
- Subject fidelity is the highest-priority constraint — the original composition, main objects, faces, and key details from the uploaded photo must remain instantly recognizable in the line drawing
- All numbers inside paintable regions and all legend text ("Instructions", "Paint the areas with the matching colors.", each color name, "{SUBJECT_NAME}") must render as legible English Latin script — this is the entire point of the artifact
- Main artwork must be pure black-and-white line-art only — no color, no gray, no shading, no texture, no gradients, no transparency, no pre-filled areas in the paintable regions; color appears ONLY in the small swatches in the legend at the bottom
- Every paintable region must be fully closed (no broken outline segments) and contain exactly one number; the numbering exactly matches the legend
- Single image output — no PDF, no multi-page layout, no colored preview version, no mockup, no contact sheet, no variant grid
- Output the image directly without explaining the prompt back
- All text in English Latin script — no substitution to other writing systems

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit photo of the subject you want as a paint-by-numbers page (without this the AI has nothing to convert — it cannot generate a useful PBN page from scratch)
- {SUBJECT_NAME} = Buddy on the Beach (the small title at the top of the page — keep it short, 2-5 words; this is what gets printed on the finished painting)
- {COLOR_COUNT} = auto-adapt between 8 and 14 colors based on image complexity (or pick one — see Bonus tips below for swaps)
- {COMPLEXITY_LEVEL} = standard (or "simple" for kids / "detailed" for adult hobby painters — see Bonus tips below)

**Generate this image:**

A single photoreal 4:5 portrait cinematic underwater photograph featuring the person from the uploaded reference image, drifting weightlessly through {UNDERWATER_SETTING}, with {LIGHT_EFFECT} filtering down through the water from above. The face, bone structure, eye shape, nose, lips, hair color, and skin tone match the uploaded reference exactly — identity preservation is the most important constraint and is locked before anything else in the scene. The person is captured mid-drift in a natural unposed moment, hair fanning slowly outward, fabric softly billowing in the current, fine bubbles trailing past skin and clothing. Light beams cut through the water in soft visible shafts, catching dust and micro-particles, casting gentle caustic patterns across skin and fabric. The mood is quietly dreamy and cinematic, like a still from a feature film — not a stock photo, not a CGI render. Subtle film grain in shadow areas, shallow depth of field with soft water haze in the distance. Single 4:5 portrait cinematic underwater image, photoreal, the uploaded person as the unmistakable hero.

**Rules the AI must follow:**
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's bone structure, eye shape, nose, lips, skin tone, and hair exactly; the person must be unmistakably recognizable as the uploaded subject
- Realistic underwater physics — hair fans softly outward not statically draped, fabric drags and clings with water resistance, bubbles trail in natural asymmetric clusters, no dry-land posing
- Realistic skin texture required — visible pores, faint freckles where natural, subsurface scattering visible under refracted light; no over-retouched porcelain skin, no glossy CGI render, no AI plastic smoothing
- Subtle film grain in shadow areas — photographic noise texture, not CG render smoothness; cinematic feature-film grading, not stock-photo flatness
- Single image output — no contact sheet, no variant grid, no before/after split, no multiple angles in one frame
- No text, captions, watermarks, or logos anywhere in the image
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, front-facing portrait photo of yourself (this is the face the AI will keep intact — without it the photo will not look like you)
- {UNDERWATER_SETTING} = a sunlit shallow tropical reef with soft coral and pale sand visible below (or pick one — see Bonus tips below for swaps)
- {LIGHT_EFFECT} = strong vertical sun shafts cutting through clear blue water with golden caustic ripples on skin and fabric (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single 4:5 portrait cinematic movie poster titled "{POSTER_TITLE}", laid out as one unified 4×3 grid of exactly 12 photoreal panels showing a full day in the life of the person from the uploaded reference image, who is a {LIFE_ROLE}. The person's face, bone structure, eye shape, nose, lips, hair, and skin tone match the uploaded reference exactly — identity preservation across all 12 panels is the most important constraint and must be locked before any scene description. The title "{POSTER_TITLE}" sits at the top of the poster in a clean cinematic film-credit-style serif, large and legible, spelled exactly as written. Below the title, the 12 panels run left-to-right top-to-bottom on a clean cinematic film-frame grid with thin matte-black dividers, each panel labeled with a small legible timestamp in the corner and an optional short caption underneath. The day arc: panel 1 (06:00) waking up, panel 2 (07:30) morning routine, panel 3 (08:30) leaving home / first commute, panel 4 (10:00) the first big task of the role, panel 5 (12:00) midday meal moment, panel 6 (14:00) a busy mid-afternoon scene, panel 7 (15:30) a small candid moment of joy, panel 8 (17:00) winding down the workday, panel 9 (18:30) evening transition, panel 10 (20:00) dinner or social moment, panel 11 (22:00) quiet wind-down, panel 12 (06:00 next day) peaceful rest. Every panel is photoreal cinematic feature-film quality with consistent color grading across the grid, subtle film grain, and the same person clearly recognizable in each one. Single 4:5 portrait poster image, 4×3 grid layout, the uploaded person as the unmistakable hero across all 12 panels.

**Rules the AI must follow:**
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt; the whole 4×3 grid lives inside this single poster
- Identity preservation is the highest-priority constraint — the same person from the uploaded reference photo must appear in all 12 panels with strict face consistency (same bone structure, eyes, nose, lips, skin tone, hair); no panel drifts to a different person
- Exactly 12 panels in a 4×3 grid — not 9, not 16, no extra mini-panels, no missing slots
- Render the poster title "{POSTER_TITLE}" and all 12 timestamps as legible English text spelled exactly as written — no Lorem Ipsum, no garbled letters, no substitution to other writing systems
- Every panel is photoreal cinematic feature-film quality — visible film grain, natural skin texture with pores and faint imperfections, no plastic AI smoothing, no cartoon stylization, no CGI render look
- Consistent color grading and lighting palette across all 12 panels so the grid reads as a unified poster, not 12 disconnected images
- Single image output — one unified 4×3 grid poster, no contact sheet of variants, no separate panels delivered as a sequence
- Output the image directly without explaining the prompt back
- All text in English Latin script

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, front-facing portrait photo of yourself (or whoever the poster is about — kid, partner, parent, friend); this is the face the AI will keep consistent across all 12 panels
- {POSTER_TITLE} = A Day in the Life of Me (or pick one — e.g. "A Day in the Life of Mom", "The Saturday of Sarah, 32", "One Tuesday in October")
- {LIFE_ROLE} = a third-grade teacher in a small-town elementary school (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single photoreal 4:5 portrait hyper-realistic cinematic fashion photograph of the person from the uploaded reference image, fully visible from head to toe, standing in {CITY_SETTING} during golden hour. The face, bone structure, eye shape, nose, lips, hair color, and skin tone match the uploaded reference exactly — identity preservation is the most important constraint and is locked before any styling or scene description. The person wears {WARDROBE}, with natural relaxed posture, confident composure, captured mid-moment as if unaware of the camera — one hand might run through hair or rest near a pocket, head slightly turned, eyes glancing off-frame for a natural candid feel, not a posed-for-camera look. The light is warm, low-angle late sun bouncing off glass buildings, casting long soft shadows across the pavement and softly rim-lighting the figure. Cinematic depth of field with the background slightly blurred but recognizable, subtle film grain in shadow areas, editorial fashion grading that reads as a magazine shoot rather than a phone snapshot. Single 4:5 portrait cinematic editorial image, golden-hour, the uploaded person as the unmistakable hero.

**Rules the AI must follow:**
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's bone structure, eye shape, nose, lips, skin tone, and hair exactly; the person must be unmistakably recognizable as the uploaded subject
- Realistic skin texture required — visible pores, faint natural imperfections, micro-asymmetry; no over-retouched magazine skin, no glossy beauty-filter smoothing, no AI plastic look
- Subtle film grain in shadow areas — photographic noise texture, not CG render smoothness; editorial-fashion color grading, not Instagram-filter saturation
- Cinematic golden-hour light from a low angle bouncing off glass / pavement — long soft shadows, warm rim light on the figure, the sun is just above the horizon
- Candid moment feel — the figure is unaware of the camera, mid-action, not a centered front-facing pose
- Single image output — no contact sheet, no variant grid, no before/after split, no multiple angles in one frame
- No text, captions, watermarks, or logos anywhere in the image
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears in the background

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, front-facing portrait photo of yourself (this is the face the AI will keep intact — without it the photo will not look like you)
- {CITY_SETTING} = a modern downtown street with reflective glass buildings and warm sunset light bouncing between them (or pick one — see Bonus tips below for swaps)
- {WARDROBE} = a relaxed-fit minimal earth-tone outfit — a cropped jacket over a fitted top, dark denim or trousers, simple leather boots (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single ultra-realistic 4:5 portrait cinematic image with high-contrast dramatic lighting and a photo-real digital-painting aesthetic, featuring the person from the uploaded reference image in a full-body side profile facing right, performing a slow graceful walking pose with powerful fashion-editorial presence. The face is copied exactly from the uploaded reference — same bone structure, same eye shape, same nose, same lips, same skin tone, same hair color and texture; identity preservation is the most important constraint and is locked before any effects or scene description. Soft wavy hair flows naturally in the wind, posture reflects quiet emotional confidence, calm sorrowful facial expression, eyes looking slightly downward, one hand gently moving forward while the other hangs naturally by their side. The person wears {WARDROBE}. Nearly half the body (the trailing left side — arm, shoulder, torso, leg, part of the hair) is dissolving into extremely thick, cloud-like grey smoke filled with glowing amber-gold dust particles; the disintegration is active, with limbs becoming transparent and breaking into luminous particles drifting back into the smoke. Behind and partially surrounding the figure is {POWER_ELEMENT}, acting as a powerful rim and backlight, dramatically illuminating the smoke, dust, and silhouette. The setting is minimal and cinematic: a desolate concrete surface beneath a heavily overcast sky. Color palette dominated by monochrome greys, deep blues, and whites, contrasted with intense amber-gold glow from the power element behind the figure. Shallow depth of field, glow/bloom lighting, particle dispersion, cinematic composition, emotional atmosphere, ultra-detailed textures, dramatic luxury-fashion energy. Single 4:5 portrait cinematic image, the uploaded person as the unmistakable hero.

**Rules the AI must follow:**
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — match the uploaded reference photo's bone structure, eye shape, nose, lips, skin tone, and hair exactly; the face stays unmistakably recognizable even as half the body dissolves
- The disintegration affects exactly the trailing side of the body (left side as the figure walks right), not the whole figure — face and front-facing half of torso remain fully solid and fully recognizable
- Natural wind movement in hair and fabric — no statically frozen 3D-render look; the figure must read as a cinematic film still, not a CGI sculpt
- Realistic skin texture on the still-solid half — visible pores, fine micro-imperfections, natural subsurface scattering under the rim light; no plastic AI smoothing
- Subtle film grain in shadow areas — photographic noise, not CG render smoothness; cinematic feature-film grading
- Single image output — no contact sheet, no variant grid, no before/after split
- No text, captions, watermarks, or logos anywhere in the image
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental signage appears

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, front-facing portrait photo of yourself (this is the face the AI will keep intact even as half the body dissolves into particles)
- {WARDROBE} = a fitted dark midnight-navy minimal outfit with clean tailored lines — long coat or pantsuit, no visible logos (or pick one — see Bonus tips below for swaps)
- {POWER_ELEMENT} = a massive glowing eclipse-like ring emitting fiery golden-white plasma light (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single 1:1 square image laid out as a 3x3 grid of nine equal panels, each panel showing the same person from the uploaded reference photo redrawn as a {STYLE_VARIANT} 3D cartoon character. Identity is the highest-priority constraint — across all nine panels the face must be unmistakably the same person from the uploaded reference: identical face shape, identical eye color and spacing, identical nose, identical hair color and style, identical skin tone, identical glasses if any in the reference, only the expression and pose change between panels. Each panel features a different reaction from the set {EXPRESSION_SET}, with the character shown shoulder-up and centered in its panel against a soft pastel gradient background unique to that panel (each panel uses a slightly different pastel color so the grid reads as a sticker pack, not a single composite). Smooth modern animation-studio 3D rendering with soft studio lighting, expressive eyes with glossy highlights, subtle subsurface scattering on the skin for warmth (not waxy), clean confident line silhouette, modern emoji-style character proportions (slightly larger head, simplified body) so each panel reads cleanly when cropped to a single sticker. Single 1:1 square image, 3x3 grid, nine equal panels, same identifiable person across all nine.

**Rules the AI must follow:**
- Aspect ratio 1:1 square — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — the face must be unmistakably the same person from the uploaded reference in panel 1 and in panel 9, no drift across the grid; preserve face shape, eyes, nose, hair, skin tone, and glasses (if any) exactly
- Exactly nine panels in a clean 3x3 grid with thin uniform gutters — no extra inset panels, no oversized hero panel, no missing panels, no 2x2 or 4x4 variants
- Each panel uses the same character with only expression and pose varying — no outfit changes, no haircut changes, no age changes between panels
- Realistic 3D-cartoon surface required — subtle subsurface scattering on skin, glossy eye highlights, natural micro-asymmetry; no over-polished waxy "AI plastic" finish, no perfectly symmetrical mannequin face
- No text anywhere — no labels under panels, no emoji captions, no watermarks, no signature
- Single image output — one 1:1 grid PNG, not nine separate images stitched after the fact
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental text appears

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit front-facing photo of yourself (or whoever the pack is for — kid, partner, friend); this is the face the AI will lock across all nine panels — without it the pack will not look like you
- {STYLE_VARIANT} = Pixar-style (or pick one — see Bonus tips below for swaps: Disney-3D / Dreamworks / modern-animation-studio)
- {EXPRESSION_SET} = happy big smile, shocked wide-eyed, slightly angry pouting, caring soft-eyed with hand on heart, confident smirk with thumbs up, sad teary-eyed, thinking finger-on-chin, laughing eyes-closed, professional polite smile

**Generate this image:**

A single 1:1 square photoreal-stylized 3D character portrait of the person from the uploaded reference image, redrawn in {STYLE_VARIANT} style. Identity preservation is the highest-priority constraint — match the uploaded reference photo's face shape, eye color and spacing, nose shape, hair color and hairstyle, skin tone, and glasses (if any) exactly; the result must be unmistakably the same person, just stylized into a 3D animated character. The character is shown shoulder-up, centered in frame, with a soft genuine closed-mouth smile and warm eyes meeting the camera, wearing {OUTFIT_VIBE}. Background is a soft pastel neutral gradient that complements the outfit. Lighting is soft three-point studio cinematic lighting with a gentle key from upper-left, fill from the right, and a subtle rim light catching the hair edge. Render quality is modern animation-studio 3D — big expressive eyes with glossy highlights, subtle subsurface scattering on the skin for warmth (skin reads soft and natural, not waxy), natural micro-asymmetry in the face (one eye very slightly higher, asymmetric smile), realistic hair strands with individual flyaways, 8K-quality render with shallow depth of field. Single 1:1 portrait, identity-locked to the uploaded reference, {STYLE_VARIANT}.

**Rules the AI must follow:**
- Aspect ratio 1:1 square — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — the result must be unmistakably the same person from the uploaded reference photo; preserve face shape, eye color and spacing, nose, hair color and style, skin tone, and glasses (if any) exactly; do not return a generic Pixar character that is only vaguely the user
- Realistic 3D-cartoon skin required — subtle subsurface scattering, natural micro-asymmetry, individual hair strands with flyaways; no over-polished waxy "AI plastic" finish, no perfectly symmetrical mannequin face, no glassy doll skin
- Single image output — one 1:1 portrait, no variants, no grid, no before/after split
- No text, captions, name labels, or watermarks anywhere in the image
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental text appears

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit front-facing photo of yourself (or whoever the portrait is for); this is the face the AI will preserve — without it the portrait will not look like you, it will look like a generic Pixar character
- {STYLE_VARIANT} = Pixar-style 3D animation (or pick one — see Bonus tips below for swaps: Disney-3D / Dreamworks / modern-animation-studio)
- {OUTFIT_VIBE} = a cozy knitted sweater in soft neutral cream (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single 16:9 landscape Studio-Ghibli-inspired storybook illustration of the person from the uploaded reference image, redrawn in Ghibli's soft hand-painted style. Identity preservation is the highest-priority constraint — the result must be unmistakably the same person from the uploaded reference, just stylized into a Ghibli painting; preserve their face shape, hair color and hairstyle, glasses (if any), skin tone, and body proportions exactly, so the scene reads as a Ghibli portrait of THEM, not a generic Ghibli character. The person is shown in the middle distance (not extreme close-up, not tiny background figure — roughly 25-35% of frame height), {ACTION}, inside {SETTING}. The world around them is filled with soft Ghibli storybook details — gentle wildflowers swaying, distant birds, soft cloud forms, hand-painted foliage with visible brushwork, painterly textures everywhere. Lighting is warm golden-hour diffuse light with long soft shadows, amber glow on hair and clothing edges, romantic nostalgic atmosphere, subtle vignette darkening the corners. Render is Ghibli-inspired wonder — photoreal glow combined with painterly illustration, soft painterly textures, hand-drawn intimacy, illustrative charm. Single 16:9 storybook scene, identity-locked to the uploaded person, Ghibli golden-hour painted style.

**Rules the AI must follow:**
- Aspect ratio 16:9 landscape — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — the person must be unmistakably the same as in the uploaded reference (face shape, hair, glasses, proportions all preserved); do not return a generic Ghibli character that is only loosely the user
- Soft painterly textures required throughout — visible brushwork, painted foliage, painted skies, hand-drawn intimacy; no flat digital clean look, no over-rendered AI plastic surfaces
- Not modern exaggerated anime — must read as Ghibli storybook illustration (Spirited Away / Totoro / Whisper of the Heart family); no sharp anime ink lines, no glossy modern-anime shading, no cel-shaded edges
- No harsh contrasts — soft natural light gradients only, no hard-cut shadows, no high-contrast digital grading
- No text, signs, captions, or watermarks anywhere in the scene — incidental world details (storefronts, books, signs) stay unreadable / decorative
- Single image output — one 16:9 landscape illustration, no contact sheet, no panels, no variants
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental text appears

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit photo of the person the scene is about (yourself / kid / partner / grandma / friend); this is the identity the AI will preserve — without it the scene will just look like a generic Ghibli illustration, not like them
- {ACTION} = walking slowly through tall grass with one hand brushing the wildflower tops (or pick one — see Bonus tips below for swaps)
- {SETTING} = a quiet sunlit meadow at golden hour with tall grass swaying and distant rolling hills (or pick one — see Bonus tips below for swaps)

**Generate this image:**

A single 4:5 portrait Disney-style storybook illustration showing the person from the uploaded reference image standing as the central character inside {SCENE_WORLD}. Identity preservation is the highest-priority constraint — the person must be unmistakably the same as in the uploaded reference photo (face shape, hair color and hairstyle, glasses, skin tone, body proportions all preserved), just redrawn in soft Disney storybook illustration style. The person is shown roughly 30-40% of frame height, standing in the lower-middle of the composition, looking slightly off-camera with a soft wonder-filled expression, wearing storybook-appropriate clothing that nods to their outfit in the reference photo (don't replace their look with a princess gown unless their reference outfit clearly suggests it). Around them, the world is filled with whimsical hand-illustrated details — ancient rune-carved trees frame a path of colorful mushrooms and glowing crystals, a mossy cottage with a stained-glass window and flower-lined balcony is nestled into the massive roots of the right-side tree which also houses shelves of tiny potions, stone animal statues rest near a stepping-stone pond. {COMPANION} float and drift around the person beneath a soft swirling pastel sky. Render style is painterly hand-illustrated Disney-storybook — visible brushwork, soft watercolor washes, hand-drawn line accents, illustrative charm; not photorealistic, not 3D-rendered. Single 4:5 portrait Disney storybook scene, identity-locked to the uploaded person, painterly storybook style.

**Rules the AI must follow:**
- Aspect ratio 4:5 portrait — strict, locked at the start and the end of the prompt
- Identity preservation is the highest-priority constraint — the person must be unmistakably the same as in the uploaded reference photo (face shape, hair, glasses, proportions all preserved); do not return a generic Disney character that is only loosely the user, do not replace their hair color or face shape with a "more Disney" version
- Painterly hand-illustrated storybook style required — visible brushwork, soft watercolor wash quality, hand-drawn line accents; NOT photorealistic, NOT 3D-rendered, NOT modern animation-studio CG
- No readable text in the scene — no shop signs, no rune-readable text, no book titles, no labels, no captions, no watermarks; world details stay decorative-unreadable
- The person is the clear hero of the composition — environment is rich but supports the person, not the other way around (the person is not a tiny figure lost in a wide landscape)
- Single image output — one 4:5 portrait, no panels, no contact sheet, no variants
- Output the image directly without explaining the prompt back
- All text in English Latin script if any incidental text appears

**Replace these placeholders with your details:**
- ⚠️ **REQUIRED — upload before pasting**: a clear, well-lit photo of the person the scene is about (yourself / kid / partner); this is the identity the AI will preserve — without it, the scene will just be a generic Disney character in a forest, not you
- {SCENE_WORLD} = a magical whimsical forest at golden hour, with sunbeams filtering through ancient trees and a soft swirling pastel sky (or pick one — see Bonus tips below for swaps)
- {COMPANION} = glowing snail spirits and sparkling light-creatures (or pick one — see Bonus tips below for swaps: pet / animal sidekick / fairy / dragon)