ONE CONTINUOUS UNBROKEN TAKE. No cuts. 20 seconds, 9:16. Handheld phone selfie, gentle micro-movements. Never locked-off. YUKI LOCK: @image1 is Yuki — photocopy her face, hair and eyes exactly as shown; wearing a pink puff-sleeve cropped top. Adult. Same close selfie framing. PRODUCT LOCK: @image2 is this exact lip gloss — same cap, tube, colours, liquid and printed text exactly as shown, never changing. Photocopy. Already in her hand from frame one. She never reaches across a desk. English, clearly audible, lip-synced. Relaxed pace. (0–5s) Looks at the lens. Speaks: “Everyone thinks getting full-looking lips means discomfort. It doesn’t.” (5–9s) Holds the gloss from @image2, casually applies it, smiles. (9–16s) Holds the gloss up to the lens. Speaks: “This one has deep hydration — locking in moisture for a smooth, volumized finish without any harshness.” (16–20s) Warm smile. Speaks: “I love it because it hydrates gently without dryness.” Background: warm ring light, blurred flowers and fairy lights. No mirror. No on-screen text. AUDIO: her English voice at the phone mic. No music. No lyrics. NEGATIVE: cuts, mute, extra people, recast face, wrong gloss bottle, overlay type, the words acid, burn, pain, sting.
Drag the clip sideways
BABY DANCE — 80s Pop @image1 is a twelve-month-old baby throwing down a razor-sharp adult-level 80s routine in neon — the running man, the robot, the cabbage patch and the sprinkler, to a bright synth-pop beat. Music only, mixed loud. One continuous 20s take, fast and joyful. Plug in ANY baby avatar.
@image1 is an absolute lock for the starting look and @image2 is an absolute lock for the revealed look: face, hair, styling, wardrobe, jewellery, materials and colours are taken from them exactly, with no drift and no reinterpretation at any point. Photographic live-action beauty-campaign rendering: soft even studio key, shallow depth of field, fine skin texture retained, neutral colour, natural motion blur on fast movement. No stylisation, no illustration, no beauty smoothing. A single locked-off centred portrait, framed head and shoulders. The camera does not move, push, pan or cut at any point, and the backdrop and framing stay identical from first frame to last. Facing the lens in the starting look, they raise both hands to the temples and run them up and back through the hair, fingers combing through to the ends, lips moving as if mouthing along to music. They lower both hands away out of the bottom of frame, the head settling square to the lens, shoulders shifting slightly, gaze steady into the camera. One hand lifts fast toward the lens, palm open and fingers spread, sweeping up across the frame until the palm fills it and the image blurs through the motion. The hand pulls away downward out of frame and clears the lens, revealing the revealed look in the same position, same framing and same backdrop. The chin lifts slightly, the head turns a few degrees, and the gaze returns to the lens, holding with small natural movement — breath, a settle of the shoulders, the curls swaying to rest — through to the end.
15-second 9:16 ultra-realistic beauty vlog. Fast 1–2s cuts, gimbal, macro application, golden morning light, LED vanity mirror. @image1 is FACE ONLY — photocopy their face. Realistic skin: visible pores, texture, peach fuzz. Do not beautify, do not de-age. Do not describe hair or clothes. @image2 is the sheet — photocopy outfit. One person only. 0–2s: Makeup-free, smile to camera, holding a brush. They SAY out loud: "Come get ready with me in under a minute!" 2–5s: Primer, foundation, concealer jump cuts. They SAY: "Keeping it light, fresh, and glowy for today's look." 5–9s: Blush, bronzer, highlighter, brows. Macro of real skin. They SAY: "A little blush and glow... because we're going for that effortless look." 9–12s: Shadow, mascara, glossy nude lip. Hair flip. They SAY: "Mascara, lip gloss... and we're basically done!" 12–15s: Walk toward camera. They SAY: "Simple, fresh, and ready for the day. What's your favorite makeup step?" No on-screen text. AUDIO (required): generate a full soundtrack. All five spoken lines must be clearly audible English lip-sync. Light upbeat music bed under the voice from 0s to 15s. Never silent. Never SFX-only.
CHARACTER LOCK — HIGHEST PRIORITY: The character in every shot is exactly the person shown in @image1, shown at miniature scale throughout — same face, facial features, hairstyle, skin tone, build and gender presentation in every shot. No identity drift between shots. Do not beautify, slim, de-age or restyle the face or hair. @image1 is identity only — face, hair, skin — not any studio backdrop, not a frozen headshot, not any specific outfit as locked wardrobe; outfit, hair and accessories are exactly whatever the uploaded photo shows, only the scale changes. Photocopy the person from @image1, shrunk to a few centimetres tall, into the porch scene below. No other people in frame. ASSET MAP: @image1 — the character. Tell the following story in exactly 3 shots, same miniature character throughout, facing an oversized everyday environment as the world of the adventure. Clear visual beginning, twist and ending — never ambiguous or faded-out. STORY: On a sunlit porch railing, the character — a few centimetres tall — stands at the edge of a gap between two large potted plants, a single strand of spider silk stretched taut across it like a tightrope, glinting in the light. A large wind chime hangs nearby, its metal tubes towering overhead and turning slowly. They step out onto the silk thread, arms out for balance, inching across the gap. A sudden gust hits, the thread whips and sways violently beneath them, and their footing gives way — but as they fall they catch hold of one of the wind chime’s dangling strings swinging past at just the right moment. The gust carries the chime string through a wide arc, swinging them clear over the gap, and they tumble safely into the leaves of the far potted plant. They lie there a moment catching their breath, then sit up, look back at the gap they nearly fell into, and break into a relieved, delighted grin. STYLE: Whimsical live-action macro realism, bright midday porch light with dappled shade from the plants, shallow depth of field, glinting silk thread and softly chiming metal tones, handheld camera with a controlled swinging tracking shot through the rescue arc. Vertical 9:16. Continuous real motion. Never freeze on the reference still. No on-screen text, no subtitles, no logos, no title card. AUDIO (native, synced, no music, no voiceover, no narration): light porch ambience and distant birdsong, a gust of wind rising, a sharp gasp as their footing gives way, the soft ring of the wind chime as they grab it, wind rushing through the swing, a soft landing in leaves, and a relieved, delighted laugh at the end. SHOT 1 (0.0-7.0s): Establish the scale clearly. Handheld macro at railing height. The character, a few centimetres tall, stands at the edge of the gap, the spider-silk thread stretched ahead of them and the towering wind chime turning slowly nearby. They step onto the thread and begin inching across, arms out for balance. The wind chime and the potted plants should read as unmistakably giant next to them — the scale contrast is the single most important thing in this shot. Face readable. SHOT 2 (7.0-15.0s): The twist. A sudden gust hits mid-crossing, the silk thread whips violently and their footing gives way. As they fall, they grab hold of one of the wind chime’s dangling strings swinging past, gripping tight as the gust carries it through a wide swinging arc over the gap below, their expression shifting from panic to startled focus. Same person as @image1, miniature scale maintained throughout. SHOT 3 (15.0-20.0s): The resolution. The swing carries them clear over the gap and they tumble safely into the leaves of the far potted plant, landing in a soft sprawl. They lie still a moment catching their breath, then sit up, look back at the gap they nearly fell into, and break into a wide, relieved grin. A clear satisfying end, not a fade-out, not a still.
CHARACTER LOCK — HIGHEST PRIORITY: The character in every shot is exactly the person shown in @image1, shown at miniature scale throughout — same face, facial features, hairstyle, skin tone, build and gender presentation in every shot. No identity drift between shots. Do not beautify, slim, de-age or restyle the face or hair. @image1 is identity only — face, hair, skin — not any studio backdrop, not a frozen headshot, not any specific outfit as locked wardrobe; outfit, hair and accessories are exactly whatever the uploaded photo shows, only the scale changes. Photocopy the person from @image1, shrunk to a few centimetres tall, into the garden scene below. No other people in frame. ASSET MAP: @image1 — the character. Tell the following story in exactly 3 shots, same miniature character throughout, facing an oversized everyday environment as the world of the adventure. Clear visual beginning, twist and ending — never ambiguous or faded-out. STORY: At dawn on a windowsill garden, the character — a few centimetres tall — picks their way between fallen leaves and blades of grass, each dew drop looming overhead like a giant glass orb. They notice a discarded bottle cap lying nearby, its curved underside catching the light like a small hull. Without warning, the first fat raindrops of a sudden shower begin to strike the ground around them like small explosions, sending up splashes taller than they are. They scramble into the bottle cap for cover just as a fast-running stream of rainwater sweeps it off the ledge, spinning and rocking through the current as sheets of rain hammer down on every side. The current slows and the cap runs aground in a shallow puddle at the base of a potted plant; they climb out onto dry ground, soaked but grinning, and duck under the shelter of a broad overhanging leaf to watch the rain fall harmlessly around them. STYLE: Whimsical live-action macro realism, overcast silver-grey storm light giving way to soft ambient glow under the leaf shelter, shallow depth of field, rain-streaked textures and splashing droplets caught in motion, gentle handheld camera energy with a fast tracking glide during the rainwater ride. Vertical 9:16. Continuous real motion. Never freeze on the reference still. No on-screen text, no subtitles, no logos, no title card. AUDIO (native, synced, no music, no voiceover, no narration): birdsong cutting off as rain begins, the growing patter of raindrops, a startled gasp as the current sweeps them off the ledge, rushing water and rain hammering around them, a soft splash as the cap runs aground, and a relieved, delighted laugh at the end. SHOT 1 (0.0-6.0s): Establish the scale clearly. Handheld macro, low angle looking up so blades of grass and dew drops tower overhead. The character, a few centimetres tall, steps between fallen leaves, notices the bottle cap resting nearby and crouches to inspect it. A dew drop or blade of grass should read as unmistakably giant next to them — the scale contrast is the single most important thing in this shot. Face readable. SHOT 2 (6.0-14.0s): The twist. The first heavy raindrops slam down around them like small explosions, splashing higher than their head. They dive into the bottle cap just as a fast stream of rainwater sweeps it off the ledge. Camera follows in a fast tracking glide as the cap spins and rocks through the rushing current, rain hammering down on every side, their expression shifting from alarm to wide-eyed thrill as they grip the rim to stay upright. Same person as @image1. SHOT 3 (14.0-20.0s): The resolution. The current eases and the cap runs aground in a shallow puddle at the base of a potted plant. They climb out onto dry ground, soaked through, catch their breath — then duck under the shelter of a broad overhanging leaf and look back out at the rain, breaking into a wide, relieved grin as it falls harmlessly around them. A clear satisfying end, not a fade-out, not a still.
30-second 9:16 ultra-premium cinematic fashion film. @image1 is the product. Photocopy that object — same olive leather grain, chunky black platform, grey laces, silver eyelets, proportions. Do not redesign. Do not turn it into a sandal or a different shoe. @image2 is the storyboard — photocopy those camera beats. Dark futuristic runway. Glossy black floor. Charcoal architecture. Haze. Olive light accents. No people, no feet, no extra products. 0–3s: Extreme wide establishing. Product on the runway. Slow mysterious open. 3–6s: Low-angle hero. Slow cinematic dolly-in. 6–10s: Smooth lateral track, then a 90° orbit. Stabilized. 10–16s: Macro of original leather texture, then close-up of dark-chrome / silver hardware. Speed ramps, match cuts. Cutting speeds up. 16–22s: Low-to-high reveal. Fast but elegant push-in. Slow-motion three-quarter rotation. 22–27s: Slow down. Final symmetrical low three-quarter hero, overhead spotlight, floor reflection. Subtle last push-in. 27–30s: Hold the hero. Fade to black. No on-screen text, no captions, no logos. AUDIO: deep cinematic bed, sub-bass, fashion-show pulse, metallic whooshes, material foley. Calm confident VO: "Some designs are made to be worn. Others are made to be remembered. Sculpted with precision. Defined by presence. This is more than footwear. This is the new icon." Music under the voice, never silent.
CHARACTER LOCK — HIGHEST PRIORITY: The character in every shot is exactly the person shown in @image1, shown at miniature scale throughout — same face, facial features, hairstyle, skin tone, build and gender presentation in every shot. No identity drift between shots. Do not beautify, slim, de-age or restyle the face or hair. @image1 is identity only — face, hair, skin — not any studio backdrop, not a frozen headshot, not any specific outfit as locked wardrobe; outfit, hair and accessories are exactly whatever the uploaded photo shows, only the scale changes. Photocopy the person from @image1, shrunk to a few centimetres tall, into the kitchen scene below. No other people in frame. ASSET MAP: @image1 — the character. Tell the following story in exactly 3 shots, same miniature character throughout, facing an oversized everyday environment as the world of the adventure. Clear visual beginning, twist and ending — never ambiguous or faded-out. STORY: On a sunlit kitchen counter, the character — a few centimetres tall — picks their way between towering utensils and a dusting of flour, a single wooden spoon lying nearby like a fallen log. A blueberry escapes from a bowl a short distance away and starts rolling steadily across the counter, straight toward the drop-off at the counter’s edge. They break into a sprint after it, weaving around scattered crumbs, gaining just enough ground to leap and land on top of the rolling berry as it picks up speed. They ride it, off balance and gripping its skin, as it rolls closer and closer to the edge — until it clips a folded kitchen towel and skids to a stop right at the lip of the counter, its momentum spent an arm’s length from the drop. They slide off, breathless, peer over the edge at the long fall below, then straighten up with a relieved, proud grin. STYLE: Whimsical live-action macro realism, warm morning kitchen light through a nearby window, shallow depth of field, visible texture on the wood counter and flour dust catching the light, energetic handheld chase camera that smooths into a controlled tracking shot once they’re riding the berry. Vertical 9:16. Continuous real motion. Never freeze on the reference still. No on-screen text, no subtitles, no logos, no title card. AUDIO (native, synced, no music, no voiceover, no narration): ambient kitchen hum, quick footsteps on the counter, a sharp gasp as they leap, the soft rumble of the berry rolling, a skidding stop against the towel, and a relieved, delighted laugh at the end. SHOT 1 (0.0-6.0s): Establish the scale clearly. Handheld macro at counter height. The character, a few centimetres tall, steps around scattered flour and crumbs, a full-sized wooden spoon and mixing bowl towering beside them. A blueberry tips out of the bowl above and drops onto the counter, starting to roll away from them toward the counter’s edge. The scale contrast — the spoon, the bowl, the edge in the distance — is the single most important thing in this shot. Face readable. SHOT 2 (6.0-13.0s): The twist. They sprint after the rolling blueberry as it gathers speed, weaving around crumbs and closing the gap. At the last possible moment they leap and land on top of it, riding its momentum as it barrels toward the counter’s edge, arms out for balance, expression shifting from panic to focus. Same person as @image1, miniature scale maintained throughout. SHOT 3 (13.0-18.0s): The resolution. The berry clips a folded kitchen towel and skids to a stop right at the edge of the counter, an arm’s length from the drop. They slide off, catch their balance, and cautiously peer over the edge at the long fall below — then straighten up, wipe their brow, and break into a wide, relieved grin. A clear satisfying end, not a fade-out, not a still.
LOOK: photographic beauty-campaign footage. Even frontal lighting on a plain white seamless, clean natural colour, realistic skin texture, shallow falloff behind the shoulders. No stylisation, no illustration, no film grain, no vignette. @image1 (bare-styling reference) is an absolute character lock for the first half and @image2 (glam-styling reference) is an absolute character lock for the second half — face, body, hair, makeup, wardrobe, materials and colours are taken from them exactly, with no drift or reinterpretation. The face is the same person in both halves; only the styling changes. @image3 (storyboard) is to be followed exactly: same framing, same hand positions, same order. CAMERA: one locked frame for the entire piece — a centred front-on bust framing, no pan, no tilt, no push, no zoom, no handheld drift. The framing, the lens, the distance and the background are identical before and after the cut. The subject stays centred and facing the lens throughout, breathing and blinking naturally. Beats, in order, each held for roughly the same time: 1. Both hands rise up into the lower frame, fingers extended, palms turning toward each other in front of the chest. 2. The palms press flat together in a single sharp clap, fingers pointing up, elbows out to the sides. 3. The palms part by a hand's width and clap together again, rhythmic and deliberate, and hold pressed together. 4. Hard cut on the held clap. The hands hold the identical position, now gloved, and the styling has changed completely. 5. The hands lower and open apart, sliding down and outward across the chest, then leave the frame at the sides. 6. Shoulders square, the chin lifts slightly and settles, holding a still pose to the lens. THE CUT: beat 3 to beat 4 is a hard instantaneous cut on the matched hand position — same height, same angle, same distance from the chest on both sides of the cut — one frame to the next, nothing in between. Do not dissolve, crossfade, morph, wipe, whip-pan or blur between the two halves. Do not animate the styling changing on screen. Do not show any intermediate state. AUDIO: sharp, close hand claps landing on beats 2, 3 and 4, with quiet room tone underneath. No spoken dialogue, no voiceover, no music. Do not generate any text, subtitles or captions. Do not generate a logo or a watermark. Do not change the background, the framing or the camera distance at any point. Do not add a second person or a mirror.
@image1 is an absolute character lock — face, body, hair, wardrobe, materials and colours are taken from it exactly, with no drift or reinterpretation in any shot. @image2 is the location. @image3 is the storyboard, to be followed exactly: same shot order, same framing, same pose and same staging in every beat. LOOK: shot on a DV tape camcorder. Soft, slightly blurry tape quality, faint tape noise, bloomed highlights under the overhead practicals, flickering auto-exposure, muted contrast, realistic skin tones. CAMERA: the subject is holding the camcorder themselves, sometimes propping it on a rack or bench. Hand shake throughout, misaligned framing, delayed focus pulls, clumsy zooms, the face occasionally cut off by the frame edge. Imperfect, unrehearsed shots. The camcorder never appears on screen. STYLE: playful and self-deprecating workout-vlog energy — exaggerated complaining, genuine laughs, exhausted but amused. Quick handheld cuts, the camera wobbling in sympathy with their unsteady legs. Beats, in order, each held for roughly the same time: 1. Propped at the rack, medium shot. They stand with hands on hips, catch their breath and let out a dramatic exhale. Dialogue: "Okay... leg day is not it today." 2. Handheld, close on the face. They wince, shake out one leg and laugh at themselves. Dialogue: "Why did I do that many reps—" 3. Propped medium shot at the machine. They sit down, push through a rep with their face straining, then flop back with a groan. Dialogue: "Okay that's it, I'm done, I'm so done." 4. Macro insert, shallow depth of field. Their hand closes around the machine handle, knuckles tensing, a faint tremor running through the fingers before the grip releases. No dialogue — ambient room sound only. 5. Handheld, walking, camera wobbling. They stand up and immediately wobble, grabbing the rack for balance, bursting into laughter. Dialogue: "Wait— okay my legs are literally shaking—" 6. Handheld, unsteady walk. They take exaggerated wobbly steps across the room toward the water fountain, the camera bouncing with their uneven gait. Dialogue, laughing: "This is so embarrassing, someone help me—" 7. Close handheld at the fountain. They lean down for a sip, sigh in relief and wipe their mouth with the back of a hand. Dialogue: "Okay... water is life right now." 8. Arm's-length framing against the wall. They lean back, still catching their breath, and grin tiredly into the lens. Dialogue: "Never doing legs again — see you guys, I'm going to go lie down somewhere." AUDIO: all dialogue spoken in English, close and slightly clipped as a camcorder's on-board microphone would capture it, breathless between lines. Ambient room tone underneath — distant plates settling, a machine cycling, faint footfall. No music. Do not generate any text, subtitles or captions. Do not generate a logo or a watermark. Do not show a camera, phone or recording device in frame. Do not add a timecode or recording indicator overlay.
Café Interview — The RegularRecreate workflow