Drag the clip sideways
CHARACTER LOCK — HIGHEST PRIORITY: the person 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. Preserve their gender presentation exactly as shown. Do not beautify, slim, de-age or restyle them. No other people in frame. Use they / their / them only. ATTACHED IMAGES ARE MANDATORY AND FINAL: The attached images outrank every word written below them — including @image2's exact shape, features and proportions. ASSET MAP: @image1 — the tiny living person. Photocopy face, hair, skin, and the clothes on the still. Ignore the grey studio pad around @image1 — that pad is not a set. Never freeze on the attached still as frame 0. Never show a checkerboard or transparency grid. @image2 — the product, exactly as shown, unchanged and not regenerated — same shape, colour, materials, hardware in every shot. No invented branding or text. There is only ONE @image2 in the entire scene, at its TRUE FULL SIZE. Never a second copy. Never a miniature replica. TINY SCALE — HIGHEST PRIORITY, MUST READ FROM FRAME 1: This is a MACRO shot. @image1 is a living person only a few centimetres tall. Their whole body would fit in a teacup. They stand or perch ON the leather of @image2 itself, working on @image2's own surface. @image2 is architecture — true full-size, same shape as the photo. It fills most of the frame like a hillside of worn leather, laces thick as ropes, hardware like metal posts. They work on its SURFACE. They are smaller than one lace eyelet. They could not reach the far end of @image2 even with a ladder of fifty of themselves. Frame 1 MUST show: tiny full-body (or head-to-waist) person standing ON @image2's own leather AND the rest of true-size @image2 filling the frame around them. If the person and @image2 look like a normal adult standing next to a normal-size object, the shot is WRONG. There is never a second @image2 at any other scale. They polish, wipe, or fix a tiny loose part ON this one object — never on a separate miniature, never on a workbench copy. FORBIDDEN CAMERA — instant fail: - Never a seated or standing adult whose body is similar in height to @image2. - Never a studio two-shot of a person beside a tabletop object of similar scale. - Never giant-hands POV / giant fingers / giant face filling the frame. - Never a socked ankle, a second foot, a second person, or a second @image2. - Never a miniature replica of @image2 anywhere in frame. If a hand appears it is THEIR tiny hand, matching their tiny body, five fingers, naturally proportioned. They are a tiny living person — not a doll, not a statue, not a posed still pasted onto the leather. Soft living micro-motion the whole time — blinks, breath in the shoulders, weight shifting. HANDS: tiny, anatomically correct, five fingers, same scale as their body. SINGLE SUBJECT: only @image1 appears as a person — no other people, figures, ankles, socks or giant hands enter frame. Only one instance of @image2 exists in the scene. SCALE REALISM: Match lighting and colour grade between @image1 and @image2 throughout so the whole shot reads as one real photograph, not a composite. The scale contrast between @image1 and @image2 is clearly visible in the very first frame. PHOTOREAL LIVE-ACTION MACRO: real camera, real light — not CGI. Match grain, colour and lighting between the tiny person and towering @image2 so the whole shot reads as one photograph. PROP CONTINUITY from frame 1: true-size @image2 filling the frame, a pin-scale cloth or tool in their hands, and the surface @image2 rests on. Nothing new appears later unexplained. CAUSALITY LOCK: every action has a shown, physical cause — they spot a smudge or mark on @image2's surface, lean toward that exact spot, and work on it; the cloth or tool catches on a ridge, stitch, or piece of hardware of @image2, causing a visible pause, not without reason; they adjust grip and continue, a direct cause and effect. PACING: real continuous motion almost the whole duration — no calm pause longer than 1 second. Shot 1 opens already in motion. Never freeze on the reference still. SCENE: a warm-lit wooden corner, @image2 at true full size filling the frame like terrain, soft lamp glow, shallow depth of field, whimsical live-action macro realism. Tiny @image1, already mid-task on @image2's own surface, is working a section of it when the shot opens. 0–2s: Already at work at insect scale — whole upper body visible on the leather of towering @image2 — they run the cloth along its surface, spotting a stubborn mark. 2–4s: The cloth catches on a ridge or piece of hardware of @image2 — they pause, steadying their footing on the leather. 4–6s: They adjust their grip and press on, working the mark carefully. 6–8s: They step back along @image2's surface to check their work from a different angle. 8–10s: Satisfied, they give the spot one last wipe until it gleams. 10–12s: They stand back on @image2's surface, small against its true full-size form, admiring the finished result. @image2 still fills the frame around them. STYLE: photoreal live-action macro, warm lamp, shallow depth of field. Vertical 9:16. Insect-scale living person on the surface of silo-sized @image2 — not a doll, not a statue, not a full-size adult. AUDIO: natural foley only — soft wiping/polishing sound, a small catch-and-pause sound, quiet resumed polishing, a satisfied exhale at the end. No music, no BGM, no voiceover, no narration. No on-screen text, no watermark, no subtitles.
Asset referencing: @image1 is the first frame of this video. It shows a hand holding THE OBJECT just above the clear water in a tall wide glass vase on a wooden table. It defines THE OBJECT, the vase, the water, the hand, the room, the lighting and the framing of the opening shot. ATTACHED IMAGES ARE MANDATORY AND FINAL: the attached first frame outranks every word written below it. Wherever the image and this text disagree, follow the image. Every colour of the pigment is a colour that already exists on @image1. THE DISSOLVE — this is the whole point of the film: the moment THE OBJECT touches the water it starts to dissolve into fine shimmering pearl pigment in its own colours. It melts away from the bottom up as it is lowered, the way a bath bomb or a sugar cube dissolves, shedding soft silky clouds of sparkling colour, until nothing solid is left. The pigment then unfolds through the water in slow silky ribbons, like liquid silk, always in exactly the colours of @image1 with a pearly sparkle, and twists into a slowly turning glowing spiral that fills the whole vase. Summary: a calm ASMR film in one unbroken take in which a hand lowers THE OBJECT into clear water, it dissolves completely into shimmering pigment of its own colours, and the colour blooms into silk and turns into a slow glowing spiral in the sunlight. Timeline: 0–2s — The framing of @image1. The hand slowly lowers THE OBJECT; its bottom touches the water and at once begins to dissolve, soft clouds of sparkling pigment in its colour pouring down from it with a gentle fizz. 2–4s — The hand keeps lowering it under the surface; THE OBJECT melts away from the bottom up in a rush of shimmering silky clouds until it is completely gone, and the empty hand opens and lifts out of the water and out of frame. 4–7s — The pigment blooms through the whole vase in slow silky ribbons and marbled swirls, curling and folding, sparkling in the sunlight. 7–10s — The swirls gather into one smooth, slowly turning glowing spiral of colour filling the vase from top to bottom, the light glinting through it. Notes: the pigment keeps exactly the colours of @image1 for the whole film. The vase stays clear plain glass. The light stays exactly as @image1 has it. The camera holds the framing of @image1 steadily, with no cuts, and the film runs as one unbroken take. No subtitles. No BGM; generate only environmental sounds and action sounds: a close intimate ASMR mix, dry and present — a soft splash as THE OBJECT meets the water, a gentle crackling fizz as it dissolves, soft bubbling as the clouds pour out, the quiet drip of water from the hand, a faint crystalline shimmer as the pigment swirls, and the quiet warm room tone of a calm old room under everything to the last frame.
A single unbroken pull-back from an extreme close-up of your face mid-scream out to a wide of the whole office, revealing you sealed inside your booth while nobody reacts. Upload your face and your booth. 18s, with sound.
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. Refer to them only as they/them in this prompt. ATTACHED IMAGES ARE MANDATORY AND FINAL: The attached images outrank every word written below them — including the product’s exact shape, features and proportions. ASSET MAP: @image1 — the tiny living person. Photocopy face, hair, skin, and the clothes on the still, exactly as photographed. Ignore the grey studio pad around @image1 — that pad is not a set. Never freeze on the attached still as frame 0. Never keep the crossed-arms studio pose. Never show a checkerboard or transparency grid. @image2 — the product, exactly as shown, unchanged and not regenerated — same shape, colour, materials, label and hardware in every shot it appears. No invented branding or text added. Ignore the grey studio pad around @image2 — that pad is not a set. Keep @image2's real label as the product's own label — do not replace that label with @image1's face. SUBJECT LOCK — HIGHEST PRIORITY: @image1 is a three-dimensional tiny living person only a few centimetres tall — same face, features and proportions, just shrunk in scale — already physically inside @image2's own real interior cavity, visible through the real material of @image2, with real volume and depth. Not a doll, not a statue, not a posed still, not a printed portrait, not a sticker, not a 2D face on the label. Soft living micro-motion the whole time — blinks, breath, weight shifting. They are inside the real hollow of @image2 itself — not inside any invented box, wrapping, capsule, shell or added packaging of any kind. @image2 keeps its own real shape, material and finish exactly as shown throughout, including after it breaks open — nothing about its appearance is invented or changed. Do not beautify, restyle or redesign either subject. FORBIDDEN CAMERA / TREATMENT: never print @image1 onto @image2's label. Never treat @image1 as label art, a mugshot on the label, or a picture glued to the product. From frame 1 the tiny person must read as a real body inside @image2. HANDS: if @image1 has hands, they are always anatomically correct with five fingers each, naturally proportioned, in every shot they appear. SINGLE SUBJECT: only @image1 appears as a person — no other people, figures or hands enter frame. PHOTOREAL LIVE-ACTION: real camera, real surface, real impact — not CGI, not an animated illustration. Match grain, colour and lighting between @image1, @image2 and the surface so the whole shot reads as one photograph. SCALE REALISM: @image1 is a few centimetres tall relative to @image2, which is at its true full size. Match lighting and colour grade throughout so the whole shot reads as one real photograph, not a composite. The scale contrast is clearly visible in the very first frame: a tiny living three-dimensional person inside the full-size product, seen through @image2. PROP CONTINUITY: the real surface @image2 is rolling along, and the edge or drop it is heading toward, are both visible from frame 1 — nothing new appears later unexplained. CAUSALITY LOCK: every event has a shown, physical cause — @image2 is already rolling when the shot opens, not starting from stillness; it falls because it reaches the edge of the surface it was rolling on, visibly; it is fragile and cracks the instant it first touches the ground, at the exact point of contact, not after a bounce or a delay; it breaks open immediately because that first-contact crack gives way at once; @image1 tumbles free only because the opening is now physically there. FORBIDDEN: a slow spiderweb crack over several seconds; staying intact after landing; bouncing then cracking later. PACING: real, continuous motion for nearly the whole duration — no calm establishing pause longer than 1 second except the brief settled beat at the very end. Each beat below must read as a distinct, clearly readable event, not a blur. Do NOT default to a calm, mostly-still shot. Shot 1 must open already in motion — @image2 is already rolling, not sitting still. Never freeze on the reference still. SCENE: a real tabletop or shelf at warm indoor light, shallow depth of field, whimsical live-action macro realism. @image2 is already rolling along the surface, @image1 a tiny three-dimensional living person visible moving inside its real interior cavity, when the shot opens. Tiny real person — not a doll, not a statue, not a picture on the label. 0–2s: Already rolling fast along the surface, @image2 approaches the edge, @image1 clearly visible as a tiny 3D body inside @image2. 2–4s: It rolls off the edge and falls, tumbling as it drops, @image1 still inside the cavity. 4–5s: The instant @image2 first touches the ground it cracks at that contact point and breaks open there — fast, easy, no delay. 5–8s: @image1 tumbles free through the opening, landing clear of the broken pieces on the ground. 8–12s: @image1 sits up or steadies themself amid the broken pieces of @image2, the full scale of both clearly visible in this final frame. STYLE: photoreal live-action macro, warm indoor light, shallow depth of field. Vertical 9:16. Tiny real three-dimensional person inside a towering full-size product — not a doll, not a statue, not a posed still, not label art. AUDIO: natural foley only — a rolling sound building speed, a whoosh during the fall, a sharp impact thud and an immediate crack as @image2 hits the ground, a tumble as @image1 comes free, a soft settling sound at the end. No music, no BGM, no voiceover, no narration. No on-screen text, no watermark, no subtitles.
ASSET MAP — every reference is bound here and cited by tag below: @image1 — the identity anchor from step 1. This is the person: it defines face, build and identity. Referred to below as @avatar. CHARACTER LOCK — HIGHEST PRIORITY: @avatar 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. 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. The grandmother described below is the ONLY other person inside the room. Every reference to the candidate in this prompt is @avatar — no pronouns, no description. ATTACHED IMAGES ARE MANDATORY AND FINAL: @image1 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 image outranks every word written below it: wherever the image and this text disagree, follow the image. SPOKEN LANGUAGE: ENGLISH. Every spoken word is in English, exactly as written in the DIALOGUE block. The grandmother speaks fluent, precise, professional English with a thick Eastern European accent. @avatar speaks native English. Deadpan office comedy, documentary realism. One locked camera, no cuts. PRIORITY 1 — THE ORDER OF THE LAST BEAT, NEVER CHANGE IT: @avatar FIRST asks the question while sitting upright and facing the grandmother, looking straight at her — and ONLY AFTER the question is finished does the searching begin. The search is the consequence of the question, never before it and never during it. PRIORITY 2 — THE SEARCH: once the question is out, @avatar searches the room for a hidden camera with big visible movements while the face stays toward the camera. @avatar leans far SIDEWAYS out of the chair to look past the monitor, tips the head back to scan the two upper ceiling corners, cranes toward the potted plant, scans along the glass wall — then snaps back upright. @avatar leans and twists but NEVER turns away from the camera; the face is visible the whole time. PRIORITY 3 — THE CONTRAST: an old village grandmother is conducting a job interview inside a sleek modern high-tech software company office. Her village clothing and the glass-and-monitor tech office are BOTH clearly visible in the same frame from the very first second. The first beat opens on this, already in progress — no establishing shot, no empty room. She is already speaking when the video starts. PRIORITY 4: the grandmother is completely serious and professional, like a veteran HR director. Calm, unhurried, authoritative. She reads from a sheet of paper and takes a note. She never smiles, never winks, never plays folksy or comic, and she never once looks up at what @avatar is doing — not even during the search. PRIORITY 5: one locked camera position, identical framing for the whole take. No camera movement, no zoom, no reframing, no second angle, no cuts. THE GRANDMOTHER: village grandmother, about 72, sturdy and solidly built, broad shoulders, calm imposing presence, deeply weathered face, black headscarf tied under the chin, dark grey hand-knitted wool cardigan over a dark floral housedress, heavy worn hands. She sits behind the desk in a modern black mesh ergonomic office chair, holding one printed sheet of paper and a cheap pen. Fully inside the frame with space around her, never cropped by the frame edge. LOCATION: small glass-walled meeting room inside a modern software company. Behind them through the glass: an open-plan office, people at standing desks, exposed ceiling, soft daylight. Inside the room: a light oak desk, a closed slim laptop, a monitor showing a plain dark screen with no text, a glass of water, a potted monstera in the corner. NO readable text, NO letters, NO logos, NO visible user interface anywhere. FRAMING: wide locked two-shot, both people centred with air on both sides. The grandmother on the left behind the desk, fully visible. @avatar on the right, seated facing her. DIALOGUE — in English, in this exact order: (0:00-0:04) GRANDMOTHER, already mid-interview and already speaking as the video opens, reading from the sheet, calm and professional: "Walk me through the difference between a monolith and microservices." (0:04-0:07) @avatar, upright, hesitant but polite: "Microservices... deploy independently." (0:07-0:11) GRANDMOTHER, writing a note, not looking up: "Good. What was your p99 latency in production?" (0:11-0:14) @avatar, still sitting perfectly upright and looking straight at her, polite and careful: "Sorry — is this a hidden camera?" (0:14-0:20) ONLY NOW the search begins: @avatar leans sideways to look past the monitor, tips the head back at the ceiling corners, cranes toward the plant, scans the glass wall — face still toward camera. Over the top of this, GRANDMOTHER, flat, never looking up from her sheet: "No. And that is not an answer." @avatar keeps searching one more beat, then settles back into the chair, defeated. (0:20 to the end of the take, only if the take runs longer than 20 seconds) Nobody speaks again. The grandmother writes one more note, still not looking up. @avatar sits still, glances once more at the ceiling corner, then back at her. Hold that silence, unchanged, to the last frame. SOUND: quiet open-plan office hum, distant keyboards, faint air conditioning. Calm, flat, conversational delivery. No music, no laugh track.
Asset referencing: @image1 is the first frame of this video. It shows two gloved hands holding THE OBJECT just above a hydro-dipping tank whose water is covered with a floating film of marble paint. It defines THE OBJECT, the tank, the marble film and its colours, the hands, the lighting and the framing of the opening shot. ATTACHED IMAGES ARE MANDATORY AND FINAL: the attached first frame outranks every word written below it. Wherever the image and this text disagree, follow the image. THE OBJECT keeps the exact shape of @image1 for the whole film. THE DIP — this is the whole point of the film: the hands lower THE OBJECT slowly and steadily down through the floating marble film into the water. The film stretches and wraps around THE OBJECT like a thin skin, clinging to every curve, carrying its swirls, curls and gold threads onto the surface exactly as they float on the water. The hands lift it out and THE OBJECT comes up wrapped in that same glossy swirling marble, water streaming off it, its shape exactly the same as before. THE WATER LINE — THE OBJECT goes into the water only up to just below its top edge, and stops there. Its top edge and every opening in it stay above the water the whole time, and the inside stays dry, clean and empty, and comes out exactly as it went in. Only the outside below the top edge takes the marble. Summary: a satisfying ASMR hydro-dipping film in one unbroken take in which gloved hands lower THE OBJECT through a floating film of marble paint and lift it out wrapped in a glossy marble pattern, dripping. Timeline: 0–3s — The framing of @image1. The hands lower THE OBJECT slowly into the film; the marble veins stretch and bend around it as it sinks, with a soft wet sound. 3–5s — THE OBJECT stops with its top edge just above the water line and holds still for a moment; the emptied film swirls on the surface around it, its opening dry above the water. 5–8s — The hands lift THE OBJECT slowly out of the water. Everything below its top edge comes up wrapped in the glossy swirling marble, water streaming off the outside and dripping from it back into the tank. 8–10s — The hands hold THE OBJECT up and turn it slowly toward the camera, the wet marble pattern shining in the light, the last drops falling. Notes: THE OBJECT keeps exactly the shape of @image1 for the whole film; only its surface changes, from plain to marble. The same two gloved hands appear throughout. The light stays exactly as @image1 has it. The camera holds the framing of @image1, tilting up gently to follow THE OBJECT as it rises, with no cuts, and the film runs as one unbroken take. No subtitles. No BGM; generate only environmental sounds and action sounds: a close intimate ASMR mix, dry and present — the soft wet sink of THE OBJECT breaking the film, gentle sloshing of water around the hands, the rush and trickle of water streaming off it as it rises, a steady patter of drips falling into the tank, and the quiet hum of a calm workshop under everything to the last frame.