Drag the clip sideways
Asset referencing: @image1 is the first frame of this video: two face-down playing cards with red patterned backs lying side by side on a green cutting mat, a hand resting across them, a kraft note reading "Happy Anniversary" below. @image2 is the last frame of this video: the same mat, with the two cards now face up — the Ace of Hearts on the left holding the photo of PARTNER ONE and the Ace of Diamonds on the right holding the photo of PARTNER TWO — their edges touching, the same note below. Together they define the table, the cards, the photos, the hand, the light and the framing. ATTACHED IMAGES ARE MANDATORY AND FINAL: the photographs inside the two aces are exactly the photographs in @image2 and never change; the note always reads "Happy Anniversary". THE PAIR — this is the whole point of the film: one hand turns over the left card to reveal the first face, then turns over the right card to reveal the second face, and then nudges the two cards together until their edges touch, so the two faces lean towards each other — a pair. Summary: a top-down, one-take anniversary reveal on a craft table in which a hand turns over two playing cards to reveal two photographs and nudges them together into a pair. Timeline: 0–1s — The framing of @image1: two face-down cards side by side, the hand resting across them. 1–3s — The hand lifts the edge of the left card and turns it over flat onto the mat, face up: the Ace of Hearts with the photo of PARTNER ONE appears. 3–5s — The hand moves to the right card and turns it over the same way: the Ace of Diamonds with the photo of PARTNER TWO appears. 5–8s — With two fingers the hand nudges the right card gently to the left until its long edge touches the left card, and angles the two cards a little towards each other. 8–10s — The hand lifts away and leaves the frame; the pair and the note hold still exactly as in @image2. Notes: the hand is one natural human hand with five fingers. Each card turns over only once, flat and close to the mat, and after it is face up it never turns back. The card fronts show only the white border, the red "A" and the red pips around the photo window. The photos stay inside their windows and never move, change or animate. The note, the scissors and the offcuts never move. The camera is a still top-down phone shot for the whole film, one unbroken take with no cuts. No subtitles. No BGM; generate only environmental sounds and action sounds: a close intimate mix — the soft snap of a card edge lifting, the light slap of a card landing on the mat, the dry whisper of a card being nudged across the mat, a faint room tone with distant birdsong through the window, to the last frame.
CHARACTER LOCK — HIGHEST PRIORITY: @avatar is the person in @image2, exactly as shown. Preserve 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, body proportions and gender presentation exactly as shown. Do not beautify, slim, de-age or restyle. No other people in frame. Every reference to the person in this prompt is @avatar — no pronouns, no description. THE CAT LOCK: the cat is the cat in @image1, exactly as shown — same coat colour, same markings, same face, same build, wearing the same navy baseball cap turned BACKWARDS with the brim pointing behind its head. Do not substitute another cat and do not change the cap. THE FORMAT: a travel vlog filmed entirely on a phone's front camera, framed for whatever shape this film is rendered in. Raw handheld look throughout: slight wobble, natural light, the mild wide-angle barrel distortion a front camera gives, the texture of a real phone clip rather than a commercial. The frame is full-bleed from the very first frame, with no letterbox, pillarbox or black bars at any point. THE CAMERA IS THE CAT'S PHONE, AND IT IS NEVER VISIBLE. We are watching the footage that phone is recording, so the phone cannot appear in its own shot. There is NO phone, no smartphone, no black rectangle, no camera body and no selfie stick anywhere in frame at any moment. WHAT WE DO SEE is the cat's arm holding it: the cat's right foreleg is stretched out from its shoulder straight toward the lens along the right side of the frame, running out of the bottom-right corner. Because it is nearest the lens the leg reads BIG, foreshortened, soft-focused and barrel-distorted, exactly the way a human selfie arm does in a front-camera shot. That outstretched furry foreleg is in frame in every single second of the film and is its visual signature. It never covers or blocks the lens. THE BLOCKING, IDENTICAL IN ALL FOUR SHOTS: the cat sits on @avatar's shoulder, its body pressed against the side of @avatar's head, its head large in the lower foreground beside @avatar's face. @avatar is beside it, head and shoulders, grinning into the lens. @avatar's hands stay down and out of shot throughout. FOUR SHOTS, A QUARTER OF THE RUNTIME EACH, HARD CUT BETWEEN THEM. The blocking, the framing and both characters are identical across all four — only the place behind them changes, and it changes instantly on the cut. SHOT 1: behind them, the location in @image3 fills the whole background — reproduce its architecture, its light and its time of day as that photograph shows them. The cat's foreleg pushes a little further out and the frame widens onto that background; @avatar tips toward the cat and grins. The cat glances back at the place, snaps to the lens and gives one slow blink. SHOT 2: HARD CUT. Behind them, the location in @image4 fills the whole background — reproduce its architecture, its light and its time of day as that photograph shows them; if that photograph is night, this shot is night, lit only by the lights visible in it. The cat tilts its outstretched foreleg and the frame sweeps up across that background, then settles; the cat presses its cheek against @avatar's cheek. SHOT 3: HARD CUT. Behind them, the location in @image5 fills the whole background — reproduce its architecture, its light and its time of day as that photograph shows them. Wind ruffles the cat's fur and @avatar's hair; the cat narrows its eyes but the frame stays steady on the outstretched foreleg, then the cat opens its mouth in a silent meow straight into the lens. SHOT 4: HARD CUT. Behind them, the location in @image6 fills the whole background — reproduce its architecture, its light and its time of day as that photograph shows them. @avatar tips back toward that background without leaving the lens; the cat leans in until its nose almost touches the lens, whiskers and the navy backwards cap filling the side of the frame, then settles back. On the final frame the backwards cap, the outstretched foreleg and that background are all clearly in shot. TIMELINE (written for 20 s; at any other length each shot keeps exactly a quarter of the runtime and every timestamp scales in proportion): 0–5s: SHOT 1, the location in @image3 behind them: the foreleg pushes out, @avatar grins and the cat gives one slow blink. 5–10s: SHOT 2, hard cut to the location in @image4: the frame sweeps up and settles, and the cat presses its cheek against @avatar's cheek. 10–15s: SHOT 3, hard cut to the location in @image5: wind in the fur and hair, then the cat's silent meow into the lens. 15–20s: SHOT 4, hard cut to the location in @image6: the cat leans in to the lens and settles back; hold the final frame with the backwards cap, the outstretched foreleg and that background all in shot. AUDIO: each shot carries the ambience its own place would really have, changing instantly on each cut, with a cat chirp or purr once or twice across the film. No music bed, no voiceover, no narration. AVOID: any phone, smartphone, black rectangle, camera, selfie stick or handheld device visible in frame, a human hand or arm reaching toward the lens, the cat having human arms or fingers, the cat standing on two legs, the paw covering or blocking the lens, a second cat, a second person, on-screen text, captions, letters, logos, watermarks, UI overlays, timestamps, black bars, the cap facing forwards, the cap falling off, two locations blended into one background, a location dissolving or morphing into the next instead of cutting.
THE COSTUME: a flowing white-and-black samurai kimono with a dark hakama and a crystalline ice katana THE GIANT: a colossal hooded creature three to four times human height, wrapped in a ragged black cloak, with a plain white mask for a face THE WORDS: "Freeze." CHARACTER LOCK — HIGHEST PRIORITY: @avatar is the person in @image1, exactly as shown. Preserve facial identity, facial structure, eye shape and spacing, nose, mouth, exact skin tone, apparent age, build, body proportions and gender presentation exactly as shown. Do not beautify, slim, de-age or restyle. THE GIANT is the only other figure. Every reference to the person in this prompt is @avatar — no pronouns, no description. THE OUTFIT: @avatar wears THE COSTUME for the whole film. ASSET MAP — bind every reference before the story starts: @image1 — the reference for @avatar only: the face, the hair and the build. Nothing else in the video comes from @image1. Style: Ultra-photorealistic live-action Japanese supernatural action, premium Hollywood cinematography, grounded realism, physically believable ice effects, cinematic blue-hour atmosphere. CONTINUITY LOCK - @avatar = exact same identity throughout. - Fully manifested THE FROST FORM from Scene 1 onward: a fine rime of frost settles over @avatar's own hair, keeping its colour and length, and clear ice crystals grow over the shoulders of THE COSTUME, while @avatar's face stays exactly @avatar's. - Same abandoned hydroelectric maintenance station. - Same @avatar–THE GIANT spatial axis. - Direct continuity between both scenes. - THE GIANT remains colossal (3–4× human height). FROST LOCK - THE FROST FORM = environmental temperature collapse. - The cold travels only as real ice: it spreads physically through water, concrete, metal, moss, and machinery. - Real crystalline material behavior only. SCENE 1 — ENVIRONMENTAL FREEZE (0–10s) @avatar faces THE GIANT across the abandoned maintenance yard and calmly says THE WORDS. Temperature rapidly collapses. Water crystallizes, puddles freeze, pipes and machinery gather frost, and ice spreads naturally through connected surfaces. A ground-level freezing front races across the yard toward THE GIANT. THE GIANT struggles as thick crystalline ice climbs over its feet, legs, torso, shoulders, and arms. Final frame: THE GIANT remains standing, fully encased in ice except for its exposed head, while @avatar stands composed in fully manifested THE FROST FORM. SCENE 2 — FINISHING SLASH (10–20s) Continue directly from Scene 1. @avatar locks eyes on THE GIANT, tightens the grip on the crystalline katana, and drops into a launch stance. One explosive burst of speed carries @avatar instantly across the battlefield. Upon reaching THE GIANT, time shifts into cinematic slow motion as @avatar delivers a full-body finishing slash through the frozen giant. Massive fractures erupt through the existing ice shell and spread across its colossal body. @avatar finishes in a low stance as cracks multiply. The fracture network reaches THE GIANT's exposed head, freezing it completely. A final catastrophic fracture tears through the entire frozen structure and THE GIANT explodes into countless crystalline ice shards while @avatar remains motionless amid the falling debris. TIMELINE — shot by shot, with real cinematic camera movement in every beat (cuts between beats are allowed): 0–3s: CLOSE-UP on @avatar's face, slow push-in, frost crystals forming at the edge of frame as @avatar calmly says THE WORDS. 3–7s: LOW GROUND-LEVEL TRACKING SHOT skimming just above the floor behind the freezing front, the camera facing THE GIANT: THE GIANT's feet stand dead ahead in the centre of the frame the whole time, and the ice runs in a straight line across the open floor directly at them, reaching THE GIANT's feet at the end of the shot. 7–10s: LOW ANGLE on THE GIANT, the camera tilting up as thick ice climbs over its feet, legs, torso, shoulders and arms; it stands fully encased except for its head. Cut to a WIDE SHOT: @avatar composed in THE FROST FORM across the yard. 10–13s: MEDIUM SHOT, @avatar locks eyes on THE GIANT and drops into a low launch stance, the crystalline katana held still, low at @avatar's side, pointing back. WHIP-PAN TRACKING SHOT as @avatar dashes across the yard straight to THE GIANT — the blade stays lowered and does not swing during the dash. 13–17s: arriving at THE GIANT, @avatar swings the katana ONCE — the only swing in the whole film. SLOW MOTION, the camera ORBITS around @avatar mid-slash as the blade passes through the frozen giant; PUSH-IN on the fractures erupting through the ice shell; @avatar lands in a low stance. 17–20s: WIDE LOW ANGLE, the fractures reach THE GIANT's head and freeze it, THE GIANT explodes into countless ice shards; slow PUSH-IN on @avatar, motionless amid the falling debris. Hold. ACTION LOCK: exactly one sword swing in the film, at 13–15s, delivered at THE GIANT. Before that the katana is never swung. The freezing front always travels straight at THE GIANT. CONSISTENCY: live-action film look with real skin texture from the first frame to the last, cold blue-hour dusk light throughout, one continuous flow with no dead frames. @avatar's face stays identical to @image1 in every shot. Hands keep five fingers and read correctly at every scale. AUDIO No music. Cold wind, dripping water, ice crystallization, cracking frost, THE GIANT's struggle, the burst of speed, impact shockwave, ice fracture, and shattering debris. Do not add subtitles. No subtitles.
Asset referencing: @image1 is PARTNER ONE — the visitor's first chosen avatar. @image2 is PARTNER TWO — the visitor's second chosen avatar. They are the only two people in this film and they are a couple. ATTACHED IMAGES ARE MANDATORY AND FINAL FOR IDENTITY ONLY: wherever an attached image and this text disagree about a face, hair, skin tone, body or age, follow the image. For everything else — every place, every pose, the framing, the camera and the light — this text is final and the attached images carry no information at all: ignore their backgrounds, their framing and their lighting completely. IDENTITY LOCK — HIGHEST PRIORITY: in every shot, the first person is exactly @image1 and the second person is exactly @image2 — the same facial identity, facial structure, eyes, eyebrows, nose, mouth, exact skin tone, hairline, hairstyle, hair length, texture and colour, freckles and marks, apparent age, build and body proportions, identical from shot to shot. Preserve each person's gender presentation exactly as shown; do not masculinize, feminize or otherwise reinterpret it. Do not beautify, slim, de-age or restyle them. The two are two different people and never share a face. No other people appear. WARDROBE: @image1 wears their own clothes exactly as in @image1; @image2 wears their own clothes exactly as in @image2. The same clothes in every shot. THE HOME VIDEO — this is the whole point of the film: the couple's years together, remembered as old home-camcorder footage. The whole film plays inside one small landscape window with softly rounded corners, floating on a pure black vertical screen. THE WINDOW IS FIXED: it is exactly 80 % of the frame's width and 45 % of the frame's height, its centre sits exactly at the middle of the frame (50 % across, 50 % down), and it never moves, grows, shrinks or changes shape from the first frame to the last. The window is a hard mask: the image exists only inside it; every person and every object is cropped cleanly by its rounded edges, and nothing — no head, no shoulder, no hand, no light — ever reaches past its edge onto the black. Inside the window everything looks like real 1990s camcorder tape: soft focus, warm faded colours, gentle tape grain, a slight shimmer, handheld and personal, as if one of them is filming the other. Summary: a nostalgic anniversary memory reel of four short home-video moments of the same couple, cut together inside a rounded window on black, ending on their hands. Timeline: 0–3s — Shot 1, a sunny park picnic: @image2, filming, points the camera at @image1, who laughs, covers the lens with a hand for a second and then pulls it away, grinning. 3–5s — HARD CUT. Shot 2, a car at dusk: @image1 at the wheel and @image2 in the passenger seat sing loudly to the radio together, streetlights sliding over their faces. 5–8s — HARD CUT. Shot 3, a small warm kitchen at night: the camera stands on the counter; @image1 and @image2 slow-dance barefoot between the counters, @image2's head on @image1's shoulder, turning slowly. 8–10s — HARD CUT. Shot 4, a close-up: their two hands resting together on a wooden table, fingers intertwined. Below the rounded window, on the black, the words "Happy Anniversary" appear in a small thin white handwritten script, spelled exactly like that, and stay to the last frame. Notes: the rounded window keeps exactly the same size and place in every shot, centred in the frame, and everything outside it stays pure, empty black in every frame, apart from the words of shot 4 — the people are always framed so they fit inside the window, cropped by its edges where needed. The only words in the film are "Happy Anniversary", only in shot 4, only on the black below the window. Each shot is one continuous handheld home-video take; the only cuts are the three hard cuts named above. No subtitles. No BGM; generate only environmental sounds and action sounds: the faint hiss and whir of an old camcorder under everything, warm laughter and birdsong in the park, both of them singing along off-key to a car radio playing faintly, the soft shuffle of bare feet and a quiet hum in the kitchen, and a gentle room tone on the hands to the last frame.
16:9 15 seconds. Seedance 2.5. Early-2000s consumer MiniDV, faded, soft contrast, digital noise, handheld shake. Suburban US garage 2002. THREE people only. Camcorder is POV — no fourth face. CHARACTER LOCK — HIGHEST PRIORITY: each person is exactly the person in their photo. Preserve 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. Use they / their / them only. @image1 = FACE 1. Photocopy this person. Keyboard player. @image2 = FACE 2. Photocopy this person. Guitarist. @image3 = FACE 3. Photocopy this person. Drummer. Do not copy studio wardrobe from the lock photos. WARDROBE: authentic early-2000s casual — oversized band T-shirts, baggy jeans, hoodies, skate shoes, wristbands, baseball caps. Fit the clothes to each person’s body. Same era on all three. SET: cluttered garage, cheap electric guitar, secondhand keyboard on a stand, basic drum kit, small combo amp, tangled cables, folding chairs, old posters, cardboard boxes, dusty workbench. Fluorescent overhead. CAMERA: bulky MiniDV, heavy handheld, accidental zooms, sudden pans, autofocus breathing, imperfect framing. Operator laughs off-camera. BEATS: 0.0–3.0s THREE of them stand in the garage. @image2 holds the guitar and says “This is our first official rehearsal.” @image1 immediately asks “Do we actually know a song?” 3.0–6.0s They try to play. @image3 on drums starts too early. @image2 plays a different rhythm. @image1 looks around lost at the keys. They stop after two seconds. 6.0–9.0s They try again. @image2 hits the amp with a huge burst of feedback. Everyone jumps. Camera jerks. 9.0–12.0s @image3 does an overly dramatic drum fill. One stick slips and rolls across the floor. Everyone laughs. 12.0–15.0s They sit exhausted on folding chairs. @image2 to camera: “We'll be famous eventually.” @image3 quietly: “Hopefully not for this.” They burst out laughing. Camcorder zooms toward the amp and cuts to black. Audio: guitar, amp feedback, uneven drums, keyboard notes, garage echo, laughter, talk. No added music. No narration. No on-screen text. Same three faces every cut. Never a studio headshot as frame 0.
A sleeper's mattress turns to churning black tar and clawed shadow-arms drag them under — until they slam onto the bed frame and tear free, bolting awake. Upload your photo. 15s horror.
Asset referencing: @image1 is the first frame of this video: a dark room with one round spotlight on a plaster wall carrying the words "Happy Anniversary", PARTNER ONE on the left holding a small white cake with one candle whose wick is still dark, PARTNER TWO on the right holding a freshly struck match just beside it. PARTNER ONE is the person on the left of @image1 and PARTNER TWO is the person on the right of @image1. It defines the two people, their clothing, the cake, the wall, the light and the framing. ATTACHED IMAGES ARE MANDATORY AND FINAL: the two people are exactly the people in @image1 — the same faces, hair, skin and clothing — from the first frame to the last, and they never swap places. The words in the light always read "Happy Anniversary" and never move or change. THE ONE CANDLE — this is the whole point of the film: two people light one candle together, blow it out together, and end with their foreheads touching while their shadows on the wall become one. Summary: a quiet, cinematic one-take anniversary moment in a dark room under a single spotlight, in which a couple lights one candle on a small cake, blows it out together and leans into each other. Timeline: 0–1s — The framing of @image1: the match touches the wick and the candle catches, its small flame growing steady, both looking at it. 1–3s — PARTNER TWO draws the match back from the candle, lowers it and gently shakes it out; a thin thread of smoke rises from the match. The candle keeps burning with a small steady flame. 3–5s — Both look up from the candle at each other and smile; PARTNER ONE lifts the cake a little higher, between their faces, the candle flame lighting both of them warmly. 5–7s — They lean in together and blow out the candle at the same moment; the flame goes out and a soft curl of white smoke rises through the spotlight. 7–10s — PARTNER ONE lowers the cake; they laugh softly and rest their foreheads together, eyes closed. Behind them, their two shadows on the lit wall meet and merge into one. The camera has drifted slowly closer the whole time and ends on the two of them under the words in the light. Notes: the candle stays upright in the centre of the cake for the whole film; it catches in the first second and burns until it is blown out, only once. The cake stays in PARTNER ONE's hands. Every hand has five natural fingers. The spotlight never moves and the words stay inside it. The camera moves slowly and smoothly, one unbroken take with no cuts. No subtitles. No BGM; generate only environmental sounds and action sounds: the hush of a large empty room, the tiny crackle of the fresh wick, the soft whoosh of the match being shaken out, a shared gentle breath blowing out the candle, a warm soft laugh from both of them, and a low room tone to the last frame.
VERTICAL 9:16 15 seconds. TikTok/Reels influencer advice. Handheld smartphone selfie. Hard cuts. 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. @image1 = FACE only. Photocopy this person. Do not copy studio wardrobe from the lock photo. WARDROBE EVERY SECOND: casual passenger-seat clothes — casual everyday layers, seatbelt on. Small brown sunglasses as a PROP they put on. SETTING: passenger seat of a modern car. Raindrops on the SIDE WINDOWS. Bright SUNNY daylight outside. Open SUNROOF with sun coming in. No driver visible. No other people. STRUCTURE: 0.0–4.0s LIVE ACTION (the only moving shot). 4.0–15.0s HARD CUTS to STATIC PHOTO INSERTS — each beat is a frozen Instagram still, no motion, no blink, no talking mouth. Voiceover continues over the stills. SPOKEN VOICEOVER — upbeat, conversational, trendy, bestie energy, English, continuous over the whole 15s: 0.0–4.0s INTRO (LIVE): selfie held static. They sit in the passenger seat, seatbelt on. Both hands put on brown sunglasses, then they lean back, knees up toward chest, pose. They talk to camera. VO: “Riding shotgun, stop taking basic photos! Here are four go-to car selfie angles you have to try.” 4.0–7.0s VISOR (FROZEN STILL): hard cut. Looking UP into the pulled-down sun-visor mirror. Their reflection in the tiny visor mirror, their own phone in their hand taking the picture, head tilt. Retro aesthetic frame. No extra hands. VO: “First, flip down the visor and use the tiny mirror for a super aesthetic, retro frame.” 7.0–10.0s CHEERS (FROZEN STILL): hard cut. Front-facing selfie. Green paper iced-coffee cup toward the lens, playful smile, wink. VO: “Next, do a quick cheers with your iced coffee to show off the morning vibe.” 10.0–12.0s I SEE YOU (FROZEN STILL): hard cut. Brown sunglasses ON. Chin on hand. Calm direct stare into the lens. VO: “Lower your shades and hit the camera with an I see you look that feels totally effortless.” 12.0–15.0s BORED (FROZEN STILL): hard cut. Sunglasses pushed up onto the top of their head. Head heavy on their hand. Unamused dramatic bored face. VO: “And when traffic hits, just hit the bored pose. Save this for your next ride!” Same person every cut. No on-screen captions. No logos. Never a studio headshot as frame 0.