15 seconds, 1:1. MiniMax H3. Cinematic live-action fused with precise black-and-white manga. Cate is ONE woman in two states. @image1 = SCENE_TRAPPED: giant open manga on the floor of the room shown in @image1. Ink Cate trapped in the page. @image2 = SCENE_BREAKTHROUGH: ink Cate tearing through paper, fingers punching out. @image3 = SCENE_EMERGENCE: REAL Cate, exactly as shown in @image3, stepping out of the torn page. @image4 = SCENE_ENCOUNTER: REAL Cate pointing at her ink self still in the book. CATE LOCK: photocopy the REAL person in @image3/@image4 — face, hair, age and outfit exactly as shown. The manga figure in @image1/@image2 is the SAME person, same outfit, ink-only. 0.0–1.0s Low wide of the giant book (@image1). Fast push-in. Ink Cate confused, looking around. 1.0–3.3s Hard cuts: face, hands searching the paper wall, profile pushing. Whisper: "Where is this? I can't get out." 3.3–4.6s Low three-quarter. Feet brace, fingers bite paper. "Open up!" Speed lines. 4.6–6.8s Paper bulges into the room. Hands, face, torso, legs convert from ink to REAL Cate (@image3) in order. Halftone becomes color. One woman only. 6.8–8.4s She lands on the floor, steadies. "I got out…?" 8.4–10.3s Smile dies. She looks back at the book. Whip pan. 10.3–11.2s @image4: REAL Cate and ink Cate point at each other. Hit-stop. Record scratch. 11.2–12.8s Smash zoom on ink Cate: "No way… you actually came out!?" 12.8–15.0s Smash zoom on REAL Cate: "Who… are you?" Brief !? graphic. Hold. Do NOT copy speech-bubble lettering from the stills except a brief ink バリッ and a final !?. No extra logos. NEGATIVE: third woman, extra people, overlay subtitles, 9:16 crop, recast hair, walking away from the book.

10 seconds, 9:16. ONE continuous LOW tracking shot. Golden hour. Photoreal. ELARA LOCK: @image1 is Elara — photocopy their FACE, hair and features exactly as in @image1. Adult. IGNORE whatever clothes are in @image1. Elara is a GIANT lying on their SIDE on a grassy hill, filling the frame like a landscape. Pose from the original film: left elbow on the grass, left hand near their cheek, looking toward camera. Right hand on a long wooden staff. Black-and-white POLKA-DOT halter crop, black choker, long black robe draped over hips and legs. NOT on their back. NOT overhead. ERIC LOCK: @image2 is Eric — photocopy their FACE. Adult. IGNORE whatever clothes are in @image2. TINY skateboarder: dark gray t-shirt, beige trousers, black sneakers, olive bag. CAMERA: extremely LOW, almost on Elara's skin. Opens on a huge thigh filling the bottom of the frame. Tracks FORWARD and slightly UP with the skater. Strong foreground motion blur. Never a bird’s-eye top-down. ERIC SKATE (real skills, curved lines, not a straight crawl): 0.0–3.0s Starts on Elara's sunlit THIGH. Carves an S-turn, knees bent, arms out, pumping the board. 3.0–6.0s Rides up onto the black ROBE at Elara's hip. Kick-turn, then a small ollie over a fabric fold, lands rolling. 6.0–8.5s Carves across the polka-dot crop toward Elara's collarbone. Weight shifts, board tilts on the curves of Elara's body. 8.5–10.0s Ends tiny on Elara's collarbone, near their face. Elara still holds the cheek pose. HOLD. AUDIO: urethane wheels, wind, grass. No speech. NEGATIVE: bird’s-eye, Elara on their back, straight-line walk without carving, extra people, overlay type, recast of @image1 or @image2.

Asset referencing: @image1 is the first frame of this video: a clean arched mirror in a wooden frame against a pale wall, a printed photo tucked into its lower right corner, one hand holding a cupcake with a lit candle, and a second hand touching the tip of a berry-red lipstick to the clean glass. @image2 is the last frame of this video: the same mirror with the words "Happy Birthday" written across it in lipstick, the same printed photo in the corner, and the same hand holding the cupcake with its lit candle. Together they define the mirror, the wall, the photo, the hands, the light and the framing. ATTACHED IMAGES ARE MANDATORY AND FINAL: the attached frames outrank every word written below them. Wherever the images and this text disagree, follow the images. THE PHOTO tucked into the frame is exactly the photo in @image2, stays in the same corner and never changes. THE GREETING — this is the whole point of the film: the lipstick writes "Happy Birthday" on the glass by hand, letter by letter, in thick waxy berry-red strokes, "Happy" on the first line and then "Birthday" on the second, each letter appearing only where the tip of the lipstick has just passed. When the last letter is finished, the writing hand lifts away and leaves the frame, and the words stay on the glass exactly as in @image2, above the steady candle flame. Summary: a warm birthday greeting in one unbroken take in which a hand writes "Happy Birthday" in lipstick on a mirror while another hand holds up a cupcake with a lit candle beside the visitor's printed photo. Timeline: 0–1s — The framing of @image1: clean glass, the lipstick tip touching it, the candle flame steady. 1–4s — The lipstick writes "Happy" across the upper part of the glass, stroke by stroke. 4–7.5s — The lipstick writes "Birthday" on the line below, stroke by stroke. 7.5–10s — The writing hand lifts away and leaves the frame; the finished words, the candle flame and the printed photo hold still as in @image2, the flame flickering softly. Notes: the letters appear only under the moving lipstick tip and never appear anywhere by themselves; once written, a letter never moves, fades or changes. The only words ever on the glass are "Happy Birthday". The hand with the cupcake stays in the same place for the whole film and the candle stays lit and upright. The mirror shows only the blank wall, never a person, a camera or a phone. The camera is handheld and nearly still, and the film runs as one unbroken take with no cuts. No subtitles. No BGM; generate only environmental sounds and action sounds: a close intimate mix — the soft waxy squeak of lipstick dragging across glass with every stroke, a tiny tap as the tip lifts between letters, the faint whisper of the candle flame, and a quiet room tone, to the last frame.

3 steps291.5K

Two dogs dance the wiggle beside a pallet of dog food, in one unbroken eight-second take. One dog opens alone, the second walks in from the right and rears up, and both finish the last beats in sync facing the lens.

5 steps

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, eyebrows, 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 otherwise reinterpret it. Do not beautify, slim, de-age or restyle them. No face changes, no identity drift. No other people share their face. ATTACHED IMAGES ARE MANDATORY AND FINAL: If an avatar image is attached, that is the person — reproduce them exactly, in every single frame, and never invent a different face, hairline, hair colour, skin tone, body or age. The attached images outrank every word written below them: wherever an attached image and this text disagree, follow the image. ASSET MAP: @image1 — the character. SCENE vs STILL: @image1 is the person — photocopy their face, hair, makeup, jewellery and clothing exactly as shown. The colourful backstage room, clothing racks, vanity bottles, oval frame and any extra figures in the still are NOT the set. The set is a dark live venue under a single warm spotlight that can die and snap back. SINGLE SUBJECT — HIGHEST PRIORITY: Only this one person is ever in frame. The audience is never shown as distinct people or faces — it exists only as implied presence through stage lighting, distant silhouette-dark space beyond the light, and crowd sound. No other individual person or hand ever appears on screen. HANDS: hands are always anatomically correct with five fingers each, naturally proportioned, in every shot they appear. Tell the following story in exactly 3 shots, same character throughout, with a clear external cause for the interruption and a clear recovery once it resolves. Clear visual beginning, twist and ending — never ambiguous or faded-out. STORY: The character is already mid-performance under a single stage spotlight, guitar in hand, singing into a microphone, fully in the moment — confident, present. Mid-note, the power cuts: the spotlight dies instantly, the amplified sound drops out, and the stage plunges into near-darkness, only a faint ambient glow left. Their voice and guitar trail off into the sudden quiet, hand stilling on the strings, eyes adjusting to the dark, a tense pause stretching as the crowd murmurs unseen around them. Then, just as abruptly, the power returns — the spotlight snaps back on, the mic crackles alive again — and instead of restarting politely from the top, they take one breath and pick the song back up exactly where it left off, quieter and rawer than before, playing it through to the end. STYLE: Live-performance realism, single warm spotlight against a dark venue that cuts to near-total darkness and snaps back, shallow depth of field, camera close on the character throughout — steady through the blackout, settling into a slightly more intimate frame once the light returns. Vertical 9:16. Continuous real motion, no freeze-frame, no cutaways to the crowd. No on-screen text, no subtitles, no watermark, no title card; any emergency or exit lighting visible in the dark stays a soft glow with no readable lettering. AUDIO (native, synced, no BGM, no added score, no voiceover): the character’s own original singing and guitar playing — an original melody, never a famous song and never copyrighted lyrics — a sudden electrical pop and hum-drop as the power cuts, their voice and strings trailing into near-silence, faint unseen crowd murmur in the dark, a tense pause, an electrical hum and mic crackle as power snaps back, one breath, then the song resuming quieter and rawer. SHOT 1 (0-7s): Open already inside the performance — no static establishing beat. Character mid-song under the spotlight, guitar in hand, singing, confident and in motion. Face and hands both readable. SHOT 2 (7-13s): The power cuts. Spotlight dies instantly, sound drops out, the stage goes near-dark. Their voice and playing trail off, hand stilling on the strings, eyes adjusting, a tense pause in the near-darkness with faint crowd murmur. The person remains visible in the faint glow — not a cut to black, not an empty frame. SHOT 3 (13-20s): The power returns. Spotlight snaps back on, mic crackles alive. One breath, and they pick the song back up exactly where they left off — quieter, rawer, more honest — playing it through. A clear, resolute ending, not a fade-out.

Asset referencing: @image1 is the first frame of this video: a dark old movie-theatre lobby, a brass-framed poster case on a red velvet wall with every bulb around it unlit, the premiere poster inside in deep shadow. @image2 is the last frame of this video: the same poster case with every bulb around the frame glowing warm and the poster brightly lit, showing THE PHOTO above the title "HAPPY BIRTHDAY" and the line "PREMIERING TONIGHT". Together they define the lobby, the poster case, the bulbs, the poster, the photo, the lettering, the carpet and rope stands, and the framing. ATTACHED IMAGES ARE MANDATORY AND FINAL: the attached frames outrank every word written below them. Wherever the images and this text disagree, follow the images. THE PHOTO on the poster and its lettering are exactly as in @image2 and never change. THE PREMIERE — this is the whole point of the film: the lights come on for the premiere. With a soft electric click, the round bulbs around the brass frame switch on one by one, running in order around the frame starting at the bottom left corner, up the left side, across the top and down the right side, until the whole frame glows. As the last bulb lights, the poster case lights up from inside and THE PHOTO and the title "HAPPY BIRTHDAY" appear bright and sharp, the warm light spilling onto the velvet wall and the red carpet. Summary: a birthday premiere in one unbroken take in which the marquee bulbs around a theatre poster switch on one by one until the poster lights up, revealing the visitor's photo above the title "HAPPY BIRTHDAY". Timeline: 0–1s — The framing of @image1: the dark lobby, every bulb unlit, the poster in shadow. 1–6s — The bulbs switch on one by one in order around the frame, each with a soft click and a warm glow, the brass catching more light with every bulb. 6–7.5s — The last bulb lights and the poster case lights up from inside; THE PHOTO and "HAPPY BIRTHDAY" appear bright and sharp. 7.5–10s — As in @image2: the whole frame glowing, the poster fully lit, the camera gliding very slowly closer, a faint warm shimmer on the bulbs. Notes: once a bulb is on it stays on for the rest of the film and never flickers off. The poster, THE PHOTO and the lettering never move, change or animate — only the light on them changes. The only words are "HAPPY BIRTHDAY" and "PREMIERING TONIGHT" on the poster. No people are in the lobby. The camera is on a smooth slow push-in the whole time, and the film runs as one unbroken take with no cuts. No subtitles. No BGM; generate only environmental sounds and action sounds: a soft electric click and a faint hum as each bulb switches on, rising into a warm steady buzz as the frame fills with light, a low deeper thunk as the poster case lights, and the distant murmur of an excited theatre crowd, to the last frame.

3 steps301.6K

20 seconds, 9:16. Macro beauty. Soft sparkly bokeh. HARD CUT between the two scenes. Photoreal 3D character. CHARACTER LOCK: @image1 is the cute 3D tub mascot — photocopy this exact character: lid, jar, label face, colours, limbs and shoes exactly as shown. Same character both scenes. Adult-safe cute. It LIP-SYNCS the English lines. 0.0–10.0s HAIR HACK. Static macro. The character stands on a coiled thick strand of brown hair. It holds a wooden spoon with white cream and golden oil. It smiles and WINKS at camera. Spoken, lip-synced: "If you mix this jelly with olive oil in equal amounts, then take a small amount and apply it to the ends of your hair, you won't need to buy expensive hair products." HARD CUT. 10.0–20.0s PERFUME HACK. Static macro. The character stands on a human wrist wearing a delicate gold bracelet. One hand holds a wooden spoon of clear gel, the other a miniature glass perfume bottle. Gentle smile. Spoken, lip-synced: "If you layer this jelly with your perfume and apply it to your skin, it helps lock in moisture and makes the scent last much longer." HOLD. AUDIO: clear spoken English, soft beauty ambience. No songs. No extra people talking. NEGATIVE: overlay type, extra mascots, walking off, different jar, silent mouth.

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, eyebrows, 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 otherwise reinterpret it. Do not beautify, slim, de-age or restyle them. No face changes, no identity drift. No other people share their face. ATTACHED IMAGES ARE MANDATORY AND FINAL: If an avatar image is attached, that is the person — reproduce them exactly, in every single frame, and never invent a different face, hairline, hair colour, skin tone, body or age. The attached images outrank every word written below them: wherever an attached image and this text disagree, follow the image. ASSET MAP: @image1 — the character. SINGLE SUBJECT — HIGHEST PRIORITY: Only this one person is ever clearly in frame. Other travelers and staff may appear only as distant, out-of-focus background movement — never a second face in focus, never another named or featured person. HANDS: hands are always anatomically correct with five fingers each, naturally proportioned, in every shot they appear. Tell the following story in exactly 3 shots, same character throughout, ending in a real, unresolved loss — no rescue, no alternate solution, no last-second save. Clear visual beginning, twist and ending — never ambiguous or faded-out. STORY: The character is already sprinting through the airport terminal, bag over one shoulder, boarding pass gripped in hand, weaving around other travelers, breathing hard, eyes locked on the gate ahead. They round the last corner just as the gate agent’s desk stands empty and the jet bridge door mechanism begins sliding shut, sealing with a heavy final clunk right as they reach it. They stop dead, hand landing flat on the closed door, chest heaving — no one to argue with, no line to join, nothing left to do. Through the terminal window beside the gate, the plane is already pulling back from the jet bridge outside. They lower their hand from the door, straighten up, and simply stand there watching it go, breath slowing, the boarding pass still in their other hand — no phone call, no new plan shown, just the moment landing fully. STYLE: Tense handheld realism building to stillness, bright terminal lighting, shallow depth of field, camera moving urgently with the sprint then settling into a static, held frame for the final stillness at the window. Vertical 9:16. Continuous real motion, no freeze-frame. No on-screen text, no subtitles, no watermark, no title card, departure boards and signage present but always soft and unreadable in the background. AUDIO (native, synced, no BGM, no voiceover, no narration): hurried footsteps and breathing, ambient terminal noise and distant announcements, the heavy mechanical clunk of the jet bridge door sealing, a sudden drop into quieter ambience, the character’s breath slowing, faint distant engine sound of the plane pulling away outside the glass. SHOT 1 (0-7s): Open already mid-sprint — no static establishing beat. Character running through the terminal, weaving past blurred background travelers, bag and boarding pass in hand, urgency visible on their face. SHOT 2 (7-13s): They round the corner to the gate just as the jet bridge door slides shut and seals with a final clunk right as they arrive. They stop hard, hand flat on the door, chest heaving — the door will not reopen. SHOT 3 (13-20s): They lower their hand, straighten, and stand still at the window watching the plane pull away outside, breath slowing, boarding pass still in hand. A clear, honest ending in the loss itself — not a fade-out, not a fix.

Asset referencing: @image1 is the first frame of this video: a closed matte-black gift box with "Happy Birthday" in silver script on its front, its lid on top, two hands resting on the lid, in a bright living room between two tall columns of black and silver balloons. @image2 is the last frame of this video: the same box open, its lid lying on the floor beside it, a cloud of black and silver balloons risen out of it, and in the middle one big clear balloon with THE PHOTO printed on it. Together they define the box, the lettering, the balloons, the photo balloon, the room, the light and the framing. ATTACHED IMAGES ARE MANDATORY AND FINAL: the attached frames outrank every word written below them. Wherever the images and this text disagree, follow the images. THE PHOTO printed on the clear balloon is exactly the photo in @image2 and never changes. THE SURPRISE — this is the whole point of the film: the hands lift the lid straight up off the box and set it down on the floor to the right of the box, and the moment the box is open a cloud of glossy black and chrome-silver helium balloons floats up out of it, their silver ribbons unrolling from inside the box. The big clear balloon with THE PHOTO printed on its front rises last, from the middle of the cloud, turns gently to face the camera and settles right above the box as in @image2, bobbing softly on its ribbon. Summary: a birthday surprise in one unbroken take in which hands lift the lid of a black "Happy Birthday" box, a cloud of black and silver balloons floats up out of it, and a big clear balloon printed with the visitor's photo rises into the middle. Timeline: 0–2s — The framing of @image1. The hands grip the lid and lift it straight up off the closed box. 2–3.5s — The hands set the lid down on the floor to the right of the box and leave the frame; the first balloons push up out of the open box. 3.5–7s — The cloud of black and silver balloons floats up out of the box, ribbons trailing, and the big clear photo balloon rises from the middle of the cloud. 7–10s — The photo balloon turns to face the camera and settles above the box as in @image2, bobbing gently, THE PHOTO clear and sharp. Notes: the box stays in the same place on the floor for the whole film and its "Happy Birthday" lettering never changes. The balloons come only out of the box and rise slowly and softly upward; none of them pops. THE PHOTO is a flat print on the balloon's surface and never becomes a person inside the balloon. The two tall balloon columns at the sides stay still on their weights. The camera stays at chest height and eases gently to the framing of @image2, and the film runs as one unbroken take with no cuts. No subtitles. No BGM; generate only environmental sounds and action sounds: the soft scrape of the cardboard lid lifting off, a light tap as it is set on the tiled floor, the rubbery squeak and soft bump of balloons brushing past each other as they float up, the whisper of unrolling ribbons, and a quiet bright room tone, to the last frame.

3 steps281.5K