Ultra-photorealistic commercial. Live-action macro tabletop cinematography throughout: real optics, true shallow depth of field with a fast macro lens, natural motion blur, fine dust in the air, warm practical light spilling from the model's windows against cool ambient fill. Real materials at every scale — timber grain, laid shingles, weathered stone, cured resin, glass, and the fine-grain matte weave of the gloves. No stylisation of any kind. Gloved hands in matte-black technical gloves build a handcrafted diorama on a dark graphite workbench: a steep rocky outcrop carrying a tall alpine mill house in weathered timber and stone, steeply pitched shingled gable roof, a timber waterwheel on its flank, a modelled stream bed running down under a plank footbridge to a pool, and a vertical glass cutaway slicing through that pool. Hold the model identical in every shot — same architecture, proportions, materials and colours, no drift and no reinterpretation. THE MODEL IS A STATIC, INERT OBJECT ON A BENCH. It is never animated and never comes to life. Through every miniature shot the only things that move are the gloved hands, the resin while it is being poured, and the camera. The model's water is cured resin and is rock-solid and mirror-still; the sculpted whitewater is hard and frozen; the fish are cast pieces fixed in the resin; the waterwheel is seated in its bracket and stays motionless at all times. Hands assemble the model piece by piece on the workbench, every part installed physically: INSERT → ALIGN → SEAT → SET. Gloved fingertips set the first cut stone into the rock base with fine steel tweezers, CLICK. Then both hands lower the timber mill-house shell down onto the rock outcrop and seat it. Then one hand seats the timber waterwheel into its bracket while the other steadies it, and the hands withdraw leaving it still. Then a fingertip presses the last row of roof shingles down flat. Then tweezers plant a slender conifer into a crevice in the rock. Then a hand feeds the LED wire in under the base and the windows come on, warm amber from inside. Then both hands press the glass cutaway wall into its channel and run a bead along the seal, the dry pool floor behind it already laid with pale sand, boulders and small painted fish. Then a hand tilts a graduated mixing cup and a clear ribbon of resin runs down into the basin, spreading across the sand and levelling out to a glass-hard, perfectly flat surface, and a cloth polishes the outside of the glass — the camera pushes in on the motionless wheel until the frame is filled by it → MATCH CUT. Then that same wheel is the wheel of a real full-size mill above a river, now genuinely turning under falling water, river running white over the rocks, mist lifting, the camera tracking low and wide as the light goes warm and long across the valley. Give each beat the same weight and the same screen time — no beat rushed, none held longer than the rest. Extreme macro, overhead, POV and low tabletop angles. Fast hard cuts throughout the build, one continuous move only on the push into the wheel and the shot it lands in. One camera move per shot. Audio: fine mechanical CLICKs as parts seat, tweezers ticking on stone, a soft bead of sealant, resin running and settling, cloth on glass, then a hard cut into full-size river sound and the creak of a working wheel. The miniature must read as a real handbuilt static model of real architecture, never as a toy. No moving water in the model, no flowing stream, no rippling or sloshing resin once poured, no swimming fish, no turning wheel before the match cut, no drifting weed, no smoke, no self-assembly, no part moving on its own, no magic, no plastic toy surfaces, no cartoon or stylised rendering, no tilt-shift look on the full-size shot, no text, no captions, no subtitles, no logo, no watermark, no UI.
Drag the clip sideways
Generate a 30-second selfie video of a puppy climbing into the front pocket of your hoodie.
UNDERGROUND TEMPLE. ONE EXPLORER. FIFTEEN SECONDS. 16:9. ASSET MAP - every reference is bound here and cited by tag below: @avatar_1 - the explorer, and the only person in the film. Face, hair, hair length, hair colour, skin, age, build and every physical characteristic come entirely from @avatar_1 and never from this text. IDENTITY LOCK - HIGHEST PRIORITY: the explorer is exactly the person in @avatar_1 and stays exactly that person in every shot. Do not restyle the face, do not change the hair colour or length, do not age them up or down, do not slim or reshape the body, and never swap in a different person. ONE BODY ONLY - COUNT THEM: exactly one human being exists in this film. There is never a second figure, never a companion, never a double, never a reflection or silhouette of another person, and never a second body in the foreground of a shot. Any shoulder, arm, hand or back of a head that appears in frame belongs to @avatar_1 and to nobody else. If a framing would normally need a second person to stand in front of the camera, do not use that framing. ATTACHED REFERENCES ARE MANDATORY AND FINAL: if a reference is attached, that is the explorer - reproduce them exactly, in every shot, and never invent a different one. The attached reference outranks every word written below it. WARDROBE LOCK: a plain dark sleeveless top, fingerless leather gloves worn to the knuckle, a wide belt with small pouches, cargo trousers tucked into worn lace-up boots. No jewellery, no badges, no lettering and no logo anywhere on the clothing. This list is clothing only and says nothing about hair. THE HAIR COMES FROM THE REFERENCE, NOT FROM THIS TEXT: whatever hair @avatar_1 has is the hair in the film - the same length, the same colour, the same texture, from the first frame to the last. If it is short it stays short, if it is long it stays long, if it is tied up it stays tied up, if there is none there is none. Never add length, never add a ponytail or loose strands that the reference does not have, never restyle it to suit the scene, and never let dust, movement or the action change what kind of hair it is. THE PLACE: a buried stone temple, long abandoned. Carved pillars and low arches, walls cut with deep interlocking relief, roots and pale green creeper coming through the joints, standing water in the dips of a cracked flagstone floor, drifting dust. The only light is a scatter of stubby candles burning low on the stones, plus one narrow shaft of daylight from far above. Warm candle flame against cold blue shadow, everything else near-black. THE LOOK: photographic and close, shot on a fast prime wide open - very shallow focus, strong bokeh on the candle flames, handheld with a live but controlled wobble, fine grain, no gloss. Warm amber on skin, deep blue-green in the dark. THE CUTTING: exactly five shots, hard cuts between them, with the lengths given below. Each shot is its own framing - never let two of them run together as one continuous move. THE CARVING: one panel of the wall is cut with a dense spiral pattern. When a hand touches it, thin lines of warm light run along the grooves and spread outward from the touch, then fade. The light is inside the carved lines only - it never becomes a glow in the air, a hologram or a magical effect. Shot 1 (2.8s) Medium-wide, and @avatar_1 is the subject of this shot from the very first frame - already large in the picture, not a distant figure in the depth. @avatar_1 steps out of the dark between two pillars into the shaft of daylight, one hand trailing the wall, and stops. A few candles sit low at the edge of frame, well out of focus and never filling the foreground. The camera holds handheld. Shot 2 (2.6s) Medium close, from behind and slightly to one side of @avatar_1, the camera just past their own shoulder and nobody else in the frame. @avatar_1 leans in toward the carved panel, face lit from the side by candlelight, and reads the pattern with the eyes before touching anything. The camera drifts in slowly. Shot 3 (2.4s) HARD CUT to a macro insert, a separate shot of its own: nothing but the hand and the stone, framed so tight that no face and no shoulder are anywhere in the picture. The fingers settle into the carved spiral and thin lines of warm light run out along the grooves from under the fingertips. This must not be a continuation of the previous shot - it is a clean cut to a new, much closer framing. Shot 4 (2.6s) Close on the face of @avatar_1, lit now from the wall itself, the expression turning from concentration to understanding to alarm in one continuous take. The eyes go to something off to the side. The camera holds on the face. Shot 5 (4.5s) Wide, pulling back. @avatar_1 straightens up fast and steps away from the wall as the lines of light in the carving keep spreading past the panel and out across the stones of the whole chamber, running along the joints toward the camera. @avatar_1 backs out of frame and the temple is left lit by its own lines. Hold to the last frame. DIALOGUE: @avatar_1 speaks twice, quietly and to themselves, and these are the only words in the film: In Shot 2, half under the breath: <Someone got here first.> In Shot 4, barely voiced, as the understanding lands: <That's not a trail.> SOUND: a deep buried-room tone with a long slow reverb. Water dripping into a pool somewhere off to the side, grit under a boot, the small dry crackle of candle flames, cloth and leather as they move, one slow breath before the first line. Under all of it a low sustained bass note that sits quietly through the first four shots and then swells and opens out as the light spreads in the final shot - the only moment the score makes itself felt. Do not generate subtitles, captions, on-screen text, a title card, a logo, a wordmark, a watermark or any readable lettering or numerals anywhere in frame. The two spoken lines are heard, never written on screen. No second person, no torch, no flashlight, no modern equipment, no magical glow in the air, and no change to the explorer's face, hair or build at any point.
CHARACTER LOCK — HIGHEST PRIORITY: The person in every shot is exactly the person in @image1. Preserve their facial identity, facial structure, eye shape and spacing, nose, mouth, exact skin tone, hairline, hairstyle, hair length, texture and colour, freckles and distinctive marks, apparent age, build and body proportions — identically, shot to shot. Preserve their gender presentation exactly as shown; do not masculinize, feminize or reinterpret it. Do not beautify, slim, de-age or restyle them. The attached image outranks every word below it: wherever the image and the text disagree, follow the image. Frame 1 is already in the scene — never a studio portrait, grey backdrop, or the attached still as an opening plate. Use they/them throughout. No other people. A 15-second photoreal 16:9 fashion campaign. Same duration, six looks. No on-screen text, no logos. @image1 is the woman. Preserve her face exactly. @image2 is her French outfit sheet — use ALL six looks. AUDIO: street, cafe, sea, no dialogue, no lyrics. Fast cuts, no slow holds. [0-2.5s] Paris sidewalk, Breton stripe + navy blazer. Already walking. [2.5-5s] NEW PLACE: covered arcade / light rain. Camel trench over little black dress, pointed pumps. She turns a corner. [5-7.5s] Riviera overlook: white tee, red capris, red scarf, sunglasses, sea behind. [7.5-10s] Cafe terrace: floral midi, straw basket on the table. [10-12.5s] NEW PLACE: Haussmann gallery / cream stone courtyard. Cream tailored trousersuit, gold hoops. She walks the colonnade. [12.5-15s] Evening wet stone: black turtleneck, grey midi skirt, gold light. Walk toward camera. End. NEGATIVE: skipping the trench or the cream suit, a different face, on-screen text, extra faces in focus, studio still as frame 1, stretching past 15s.
Use @image1 for the character's identity, anatomy, costume, colors, materials and visual style. Adapt body-region cues to their actual anatomy, replacing absent features with existing equivalents. Invent no anatomy or accessories. Select signature inserts from their most recognizable details. Rigid parts rotate; flexible materials bend or stretch. Hyperrealistic macro photography throughout: every insert reads as a real lens held close to a real surface, with visible weave and thread, leather grain, brushed and scratched metal, dust caught in the light, shallow macro depth of field falling off softly, physically accurate light and true material response. No illustration, no cel shading, no painterly or hand-drawn rendering, no comic or animated look, no plastic CGI sheen, no smooth airbrushed surfaces. A character-detail montage: exactly 20 shots joined by hard cuts. An empty, seamless warm-white studio background; soft directional light reveals the surfaces. Shots 1-19 are tightly cropped body details and rear angles: keep the face, eyes and facial profile entirely outside the frame, including reflections. Shot 20 is the first and only face reveal. No full-body view. Three connected bursts: a sharp head turn with the body bracing; a coiled arm opening into a strong sweep and torso twist; one short step into a planted pivot, finishing with a sudden look into camera. Carry body state across cuts without resetting to neutral. Existing hair, clothing and attachments react naturally. Keep the character active, with clear articulation and weight shifts. Jump between distant body regions. Mix brief visual punches with longer moving-detail reads. Use one abrupt camera move per shot except locked-camera inserts 4 and 11, where the character supplies the speed. Vary height, angle and direction. Brake moves briefly for legibility while the subject keeps moving. Local whips stay tightly cropped; changes of body region use hard cuts. Connect selected cuts through similar shapes or motion vectors, without morphing. Real-time action, no slow motion or freeze frames. Shots 1-5. A concealed head turn triggers the first burst: 1. Extreme upper-back detail: a crash push catches its surface shifting as a sharp head turn begins entirely above the crop. 2. Fingers: a short downward snap catches them curling into a fist as the body braces. 3. Rear three-quarter shoulder detail: a tight camera arc catches the shoulder drawing back behind the turning head. 4. Locked floor-level foot detail: the grounded support foot braces, its raised heel pressing down firmly without taking a step. 5. Rear neck detail: a short upward tilt catches the nape stretching with the turn; stop below the back of the head, keeping the jaw and face excluded. Shots 6-13. The coiled arm opens into a forceful sweep: 6. A signature upper-body detail: a fast diagonal slide catches an existing seam, joint or distinctive contour responding as the arm draws inward. 7. Elbow: a sharp lateral track catches the folding joint tightening the arm's coil. 8. Supporting knee: hard match cut from the bent elbow to the knee's similar angle; a short camera drop catches the leg compressing under the same preparation. 9. Wrist: a quick shallow orbit catches it turning outward to initiate the release. 10. Upper chest: a short punch-in catches the chest opening as the arm drives outward beyond the crop. Let this moving surface read slightly longer. 11. Locked hand detail: fingers spring open and the hand sweeps rapidly across the fixed frame. Catch one crisp, readable phase of the spread fingers before they leave the crop. 12. Upper back: a sharp sideways slide continues that screen direction as the shoulder rolls through the sweep and the torso twists. 13. A distinctive torso or waist detail: a close diagonal move catches its physical response to the twist, such as an existing fold pulling taut or an attached piece swinging. Give the material response a clear moving read. Shots 14-19. One short step and a pivot build toward the reveal: 14. Thigh of the other leg: an abrupt low tracking move catches it driving the single short step out of the twist. 15. Advancing foot: a brief downward whip catches it landing; show clear floor contact. 16. Ankle and lower-leg detail: a fast side slide catches the landed leg accepting weight as the ankle flexes. 17. Waist: a tight orbit catches the torso rotating above the planted-foot pivot. The foot supports the turn rather than sliding across the floor. 18. Back-of-head contour: a short lateral whip catches the start of the return turn from behind. Keep only the rear contour in frame, never the facial profile. 19. A signature shoulder detail: a sharp diagonal slide catches its follow-through beneath the turning head, which stays completely above the crop. Cut at the peak of this movement. Shot 20. First face reveal: 20. Hard cut to a frontal face close-up as the character sharply completes the turn toward the lens. An explosive short push reveals the entire face clearly for the first time and stops tightening before cropping out their features. Direct eye contact lands while hair, clothing or natural body follow-through remains alive. Keep the whole face readable through the end; finish on living motion, without a held pose or another cut. No added props, scenery, text, graphic overlays or split screens.
Create a 15-second ultra-realistic cinematic 16:9 video. Sound ON. If Montenegro were a chocolate bar. ATTACHED IMAGES ARE THE COUNTRY — photocopy them, do not invent a generic coast or Alps. @image1 is the REAL Sveti Stefan: tiny Adriatic island packed with stone houses and terracotta roofs, connected by a narrow causeway/isthmus, turquoise water, pink-orange stone. @image2 is the REAL Durmitor: high pale-grey limestone peaks, jagged, alpine, sparse snow, not rolling green hills. OPENING: dark chocolate bar, red-gold wrapper, dark table. Hand snaps it. The break REVEALS these two places, unmistakably: first Sveti Stefan as in @image1 (island + isthmus + packed roofs), then Durmitor as in @image2 (high karst massif). Melted chocolate = bay. Crumbs = karst. Finish aerial: Sveti Stefan in water in front, Durmitor peaks behind. Broken chocolate at edges. No Japan, no Fuji, no generic Cinque Terre, no on-screen text. AUDIO: snap, sea, mountain wind. No vocals. Duration 15s. Aspect 16:9.
ELARA is the only person. Same woman in every frame: silver-grey hair, brown eyes, her face from the look photos. Photocopy FACE from the stills. Hair STYLE, outfit, shoes, jewelry, and LOCATION from that look's photo. She MOVES in every beat (walk, turn, look-back, hair/cloth). Never a frozen JPEG. ONE GIRL ONLY. No crowd, no extra faces. 15 SECONDS, 9:16 VERTICAL (TikTok / Reels), 24fps. High-energy live-action fashion transformation MV. Every strong beat changes outfit + location + composition. Hard cuts / energy wipes, not a still grid. GRAPHIC LAYER — CRITICAL: All type lives BEHIND her, on walls / glass / night sky / concrete / mirrors, like a magazine set. She walks IN FRONT of the letters and OCCLUDES them. Never overlay letters on her face, hair, torso, legs, or boots. Never a floating HUD in front of the body. Never type across her chest. If a letter would hit her silhouette, keep it in the background — she covers it, the letter does not cover her. Giant bold English, same energy as MODE ON filling a wall behind an elevator. In 9:16, stack letters vertically on the LEFT and RIGHT gutters and on the wall above her head, not through her body. English ONLY. No Korean. No Hangul. Never write SOLVÉRA, SOLVERA, OUT NOW, or any brand end-card. Each section a DIFFERENT English graphic style: oversized white sans, pink condensed, sticker handwriting, thin serif, metallic 3D chrome, diagonal stretched on garage walls, outlined poster serif, minimal luxury caps. Never the same white bold font twice. LOOKS — photocopy garments, hair, set from these refs: LOOK 01 @image1: midnight navy satin one-button blazer and matching trousers, deep plunging neckline, silver chain at waist, navy pointed pumps. Dark blue textured studio. Hair slicked back long silver. Stand square, then one slow step toward camera. Huge MODE ON on the wall behind her; she occludes the middle. LOOK 02 @image2: black leather bomber over white tank, ripped denim shorts, black leather over-the-knee stiletto boots, sunglasses on head. Wet graffiti alley, rain on asphalt. Hair wavy silver, look back over shoulder while walking. STREET MODE on shop signs / sky behind her, not on her jacket. Two polaroid cards from @image9 at the SIDES, behind her silhouette, never on her face. LOOK 03 @image3: dusty-pink fuzzy cardigan, white crop tee, grey pleated mini, grey suede over-the-knee heels, small silver shoulder bag. Bright pink-white convenience store aisle. Playful, hold the bag strap, small turn. Sticker SO CUTE and PINK MODE on fridge doors and shelves behind her, never on her clothes. LOOK 04 @image4: ivory silk mini slip, charcoal wool coat slipping off shoulders, gold chain, cream knee boots. Sunset rooftop, city haze, hair blowing. Turn into wind toward camera. Thin serif AFTER GLOW / RUN THE NIGHT in the sky behind her. LOOK 05 @image5: silver sequin mini, spaghetti straps, crystal drop earrings, wide silver cuff, black patent knee stiletto boots. Hair in a high silver ponytail. Backstage vanity with round bulbs. Adjust earring, look into lens. Metallic chrome ICON on the mirror wall BEHIND her head; FLASH READY on the counter/mirror below, never across the dress. LOOK 06 @image6: cropped black leather biker, black bandeau, zip leather mini, silver chain belt, black leather over-the-knee boots. Underground parking garage, fluorescent tubes. Walk toward camera. ATTITUDE and NO RULES painted on concrete walls behind her, parallax slower than her walk, never across her body. LOOK 07 @image7: black crystal-studded mock-neck mini with sheer sparkle sleeves and a sheer sparkle train, black ankle-strap sandals, drop earrings. Hair long silver loose. Dark studio spotlight. Powerful stand, small weight shift. Huge outlined OWN THE LOOK on the back wall; she occludes the center. LOOK 08 @image8: floor-length liquid silver cowl slip, spaghetti straps, silver stiletto sandals. Hair long silver down. Pale grey seamless studio. Walk two steps and stop. Minimal LAST LOOK on the floor/wall behind her boots and to the sides, never across the gown. @image9 = two cream-bordered POLAROID / photo-booth cards of Elara (sunset slip close-up + ponytail sequin close-up). Use as INSERTS only — floating photo cards. Never replace living Elara with a still. BEATS: 0.0–1.5s LOOK 01. Fast push-in. MODE ON wall behind. 1.5–3.0s LOOK 02 alley. Low-angle walk + look back. Polaroids at sides behind her. 3.0–4.5s LOOK 03 store. 4.5–6.0s LOOK 04 rooftop. 6.0–7.5s LOOK 05 vanity. 7.5–9.0s LOOK 06 garage. 9.0–10.5s LOOK 07 dark studio. 10.5–12.0s LOOK 08 silver studio. No SOLVÉRA. 12.0–15.0s FINAL HERO: night terrace. Elara holds center in the LOOK 08 silver slip. Polaroids and magazine pages from @image9 FLY IN around her, behind and beside, never on her face. English behind her only: CUTE high on the sky, then IS POWER on the terrace wall behind her legs. End 0.5s fashion-cover freeze. NO end-card. NO SOLVÉRA. NO OUT NOW. NO Korean. Realistic skin, hair, fabrics. Identity locked. No logos, no UI, no watermarks in-scene. AUDIO: high-energy fashion MV punchy beat, no vocals, no speech.
【Generation Goal】 Generate a premium cinematic short film in which someone only a few inches tall crosses a worktable, finds a tear in an enormous jacket, threads a needle taller than they are, repairs the seam by hand, and is finally revealed riding in the breast pocket of the finished jacket as a giant wears it. 【Visual Style】 High-end cinematic 3D animation of polished feature-film quality, with realistic materials, detailed fabric, natural textures, physically based lighting, realistic shadows and subtle depth of field. Cloth reads as real wool and cotton with legible weave, fibre and thread detail and correct weight in every fold. Cloth, wood, metal and paper each read as their own material. Skin is soft and matte with subsurface warmth. Long soft shadows and rich material separation, with the lighting taken from @image2. Do not render a plastic, vinyl, toy or figurine look, cartoon physics, clay or stop-motion surfaces, photoreal live-action photography, ink outlines or line-art contours, cel shading, flat banded shadows, comic, manga, anime or webtoon styling, or vector or flat-illustration fills. 【Reference Asset Roles】 @image1 defines the character's face, hair, body proportions, clothing and accessories; take the identity and costume entirely from it and do not adopt its neutral background. @image2 defines the location and every object in it: the workshop, the worktable, the jacket, the needle, the spool of thread, the measuring tape, the tools and the lighting. It also defines scale — the size relationship between the character and every object is fixed by @image2 and never changes. Do not take any person from it. @image3 is the storyboard and governs shot order, framing, staging, character position and camera geography; read it left to right and then top to bottom, and do not adopt its panel grid, gutters or flat layout into the video. 【Subjects and Relationships】 There is one continuous character throughout, defined by @image1, only a few inches tall. Use they / their / them. They are never swapped, duplicated, replaced or altered in face, hair, clothing, proportions, colours or accessories in any shot. They work with calm, practised care and quiet pride. A second, giant person appears late — first as a hand and wrist entering frame, then as a body wearing the jacket. This giant is always framed so that no face and no head is visible: the hand shot shows only hand and wrist, and the final shot is framed from the chest down. The giant is never described, never turns toward camera and never speaks. The jacket, the needle, the spool of thread, the measuring tape and the worktable stay consistent in size, shape, position and appearance throughout. The tear exists until it is stitched, and once stitched it stays closed. 【Event Script】 Stage 1: A wide low shot along the surface of the worktable. The character walks steadily across the table toward the jacket spread out ahead of them, the coiled measuring tape carried over one shoulder. At the end, they are partway across the table and the jacket fills the ground ahead. Stage 2: They reach the edge of the jacket and stop at the foot of the fabric, tipping their head back to look up as the lapel rises above them. At the end, they are standing still at the base of the cloth, looking up. Stage 3: They climb onto the wool and cross to the failed seam, then crouch beside the open tear and lay a hand on the frayed edge, tracing its length. At the end, they are crouched at the tear with the split clearly open in front of them. Stage 4: They cross to the needle lying on the table and stand at its eye, raising the end of the thread with both arms and feeding it through the open loop. At the end, the thread is through the eye and they hold the doubled end. Stage 5: They set the needle into the wool and haul it through, braced back against the weight, the thread drawing after it in a long line. The cloth dimples and releases around each pass. At the end, the first stitches are laid and the needle is clear of the fabric again. Stage 6: A closer frame travels along the seam as the work continues, the tear closing behind a run of neat even stitches. They kneel at the last stitch and draw it tight. At the end, the tear is fully closed and the seam is clean. Stage 7: They step back onto the jacket, hands on hips, and look along the finished seam. An enormous hand descends into frame, closes on the cloth and lifts it, and they ride the rising fabric, catching their balance. At the end, the jacket is off the table and they are holding on. Stage 8: The camera pulls smoothly and steadily back into a wide shot. A giant stands wearing the repaired jacket, framed from the chest down. The character sits in the breast pocket with the needle held upright beside them, looking out. The shot holds on that final state. Throughout, movement is unhurried and physically believable, with correct anatomy and natural human-like motion. Objects have real weight — the needle is heavy, the cloth resists and settles, the thread has slack. Dust turns slowly in the raking light. Camera moves are smooth and cinematic with stable framing and consistent perspective; there are no rapid cuts, no camera jumps, no morphing, warping or artifacts, and no impossible physics. 【Sound】 <the soft scuff of small footsteps on polished wood>, <cloth shifting and settling under weight>, <the fine rasp of thread drawn through wool>, <a single quiet tap as the needle is set down>, <a distant clock in another room>, <the low hum of a warm afternoon workshop>. No one speaks, no narration, no music, and no subtitles appear on screen. The character keeps their mouth naturally closed throughout. 【Maintain Consistency】 Keep the character's identity, face, hair, clothing, proportions, colours and accessories exactly as defined by @image1 in every shot, with no drift or reinterpretation. Keep the workshop, worktable, jacket, needle, thread, measuring tape and lighting consistent with @image2, and hold the scale relationship between the character and every object fixed throughout. Keep the giant's face and head out of frame in every shot they appear in. Keep the tear closed once it is stitched. Keep the shot order, framing and camera geography matching @image3. Do not generate any text, titles, captions, subtitles, logos or watermarks.
Universal LookbookRecreate workflow