Amateur behind-the-scenes phone video, ~6s, low at the concrete tank lip beside the dolly track, inside the ENORMOUS cavernous aircraft-hangar-sized soundstage — the same tiny model of a dense island metropolis of packed rectangular skyscrapers with a tall stepped art-deco spire, a needle-topped tower and a long suspension bridge in the MASSIVE in-floor water tank, the COLOSSAL blue screen with orange crosses towering off the top of frame, crew at the hazard-striped tank lip, camera crane + steel dolly-track. Handheld, ultra-wide phone lens, sensor grain, auto-exposure clipping the LED lights. 0–2s ESTABLISH — the huge quiet set holds; the tiny model is calm, full-size crew wait beside the hazard-striped tank lip, the cavernous room and towering blue screen dwarf the miniature while the practical creature rig hangs just out of frame. 2–4s TRIGGER — a director calls "Action!" off-camera; the giant practical creature rig activates as hydraulic mechanisms drive an enormous original monster foot downward toward the miniature. 4–6.5s IMPACT (hero beat) — the giant practical creature foot crashes down through the model, storeys tall against the tiny city, crushing buildings into real dust plumes and flung debris, a shockwave of grit rolling outward; crew flinch and step back, the crane swoops, chunks of miniature masonry and dust blast past the phone camera, the towering blue screen remains dominant with the debris plume rising many storeys high. 6.5–8s AFTERMATH — dust hangs in the huge studio volume, rubble settles across the flooded miniature streets, water sheets off the tank rim, the handheld camera re-settles, loop-ready. Sound: Diegetic only — a director's off-camera "Action!" call at the trigger + hydraulic machine rumble, heavy impact thud, cracking miniature structures, crew shouts, rig motors, water splashes and lingering debris fall; the impact roar arrives exactly with the stomp; no music, no narrator. Amateur behind-the-scenes phone photo inside an ENORMOUS cavernous film soundstage the size of an aircraft hangar. A tiny handbuilt MODEL of a dense island metropolis of packed rectangular skyscrapers with a tall stepped art-deco spire, a needle-topped tower and a long suspension bridge sits at model-railway scale on a raised shoreline inside a MASSIVE in-floor water tank sunk flush into the concrete floor (no table, no legs), ~30m across, grey hazard-striped coping and coiled cables round the rim. Towering behind and above: a COLOSSAL BLUE chroma-key wall rising ~5 storeys and running off the top of the frame, covered in many small orange "+" tracking crosses in a dense grid. A massive practical giant original creature stomp tears across the model and erupts MANY STOREYS HIGH, dust and debris cresting far up the screen with empty air still above. Full-size crew in black "EFFECTS CREW" and grey "SPECIAL EFFECTS" tees at the tank lip stand as TALL as the model towers, a distant row tiny against the far wall; camera crane/jib + steel dolly-track in shot. Wide phone framing — model in the LOWER THIRD, towering screen + air filling the top two-thirds, nothing fully contained. Flat cool LED wash, sensor grain, mild handheld, auto-exposure clipping the dust — a real phone clip, NOT a render. A giant practical creature foot crashes down through the model, storeys tall against the tiny city, crushing buildings into real dust plumes and flung debris, a shockwave of grit rolling outward; a physical monster rig on set, not a render. Negatives: NOT a tabletop diorama / desk model / model on a table with legs; NOT a cramped small room or low ceiling; NOT the model filling the frame or close tight framing; NOT a short blue screen with its top/side edges in frame; NOT a few large sparse crosses; NOT the whole room contained; no CGI/3D render/game-engine/videogame, no cartoon/anime/illustration, no glossy Octane sheen, no plastic/waxy, no HDR overcook, no fantasy glow, no clean studio-photo look; stays a small MODEL in a huge studio, never the real full-size city; no captions/subtitles/on-screen text, no watermark, no logos or brand wordmarks, no aspect ratio, no AI tool names.
School Rooftop Confrontation
A cinematic slice-of-life video captures a female student holding her backpack, surrounded by several figures on a school rooftop. The scene escalates into an action sequence with her executing close-range punches, elbow strikes, and sweeping kicks with sharp, decisive movements.
Prompt
【画质风格】真人实拍级韩国电影写实质感,全程手持微晃镜头叠加院线级冷调色,青灰偏蓝的阴天色温,肤色保留冷白通透;针织开衫的绒感、百褶裙摆的翻动、发丝飘散、雨前阴云的压抑层次全部写实还原;粗粝胶片颗粒,浅景深人物特写为主;8K 电影级拍摄质感,2.39:1 宽画幅。【角色设定】 · @她(女主,图片1):按参考图还原——黑发渐变粉色中长卷发,珍珠耳饰,白衬衫系酒红色蝴蝶结领结,米色针织开衫,黑色百褶短裙,黑色中筒袜配黑色乐福鞋;眼神从隐忍低垂到抬眼锐利,动作干脆利落,发力刚猛,擅长近身直拳、肘击与利落的扫踢。 · 围堵人群:五六名穿深色制服风衣着的高个身影,面目始终处于虚焦或背光中,不给正面特写,只作为压迫感群像存在。 · 场景:阴天放学后的教学楼天台,水泥地面,四周是锈迹斑驳的绿色铁丝网围栏,远处城市天际线灰蒙,风把云压得很低,地面散落着被扔掉的书包和课本。【运镜逻辑】开场手持镜头从她低垂的脸缓推,建立压抑;书包落地为爆发信号,镜头骤然一沉切入极速;随即进入六连切快速蒙太奇,每一切都是反击特写,快切之间插入两次不到一秒的微慢动作(Micro Slow-mo)拉出"快慢快慢"的爽剧节奏;群体溃退后镜头绕至她背后仰拍全景;最后一击时间放慢,镜头绕侧定格;收尾镜头缓缓后拉,全程张弛交替。【完整画面流程】开场:手持镜头微晃,天台上,她被五六道深色身影围在铁丝网前,低着头,粉色发梢被风吹得贴在颊侧,酒红色领结微微歪斜。一只手从画面外伸来,狠狠拍掉她怀里的书包——书包砸在水泥地上,课本散开的一瞬,她垂着的眼睫抬起,眼神骤然锐利。爆发:她贴着最近一道身影欺身而进,不退反进的瞬间镜头一沉,画面切入六连切快速蒙太奇,快慢交替,节奏抓人: 1. 第一切【快】:近身直拳特写,她拧腰送肩一记直拳打在迎面身影胸口,对方整个人向后踉跄撞进同伴怀里,两人一起跌退,她的粉发梢甩出一道弧线 2. 第二切【微慢】:抓腕反制特写,侧后方一只手抓向她肩膀,画面骤慢——她反手扣住对方手腕拧转,针织开衫的袖口绷紧,对方身体被带得侧翻,悬在半空的瞬间发丝与裙摆同时展开,随即画面弹回极速,对方摔在水泥地上滑出半米 3. 第三切【快】:低位扫踢特写,她顺势沉身一记扫堂腿,乐福鞋鞋跟贴地划出半圆,正面扑来的身影被扫得双脚离地、横着摔倒,黑色中筒袜绷直的小腿线条利落有力 4. 第四切【快】:肘击连段特写,两道身影同时从左右合围,她左肘后顶、右肘横扫,两声闷响连成一线,左右身影各自捂着腹部弯腰跪地,她站在两人中间,裙摆还在惯性摆动 5. 第五切【微慢】:躲拳反击特写,最高的一道身影抡起拳头当头砸下,画面再度骤慢——她只侧头半寸,拳风擦着粉色发梢掠过,几缕断发缓缓飘起,她的眼睛在慢镜里冷冷抬起看向对方;画面弹回极速,她一记膝撞顶在对方腹部,对方弓身抱腹跪倒 6. 第六切【快】:拽领终结特写,她一把拽住最后一人的衣领往下带,对方重心前栽的同时她抬膝再进半步急停——膝盖停在距对方面门一寸处,风压掀起对方额发,对方吓得瘫坐在地连连后爬六次切镜一气呵成,快慢两档交替咬合,每一帧都拉满反击张力。收束:镜头绕至她背后低角度仰拍全景,五六道身影或坐或爬,连滚带爬地退向天台门,门"哐"一声撞开又弹回。风掀起她的裙摆和开衫下摆,粉色发尾狂舞,她站在原地缓缓松开攥紧的拳,指节还微微发颤。终镜:画面放慢,她转过身,弯腰捡起地上的书包,一本一本把散落的课本拾回怀里,拍掉封面的灰。起身时她理了理歪掉的酒红色领结,对着镜头方向抬起眼——眼神已经恢复平静,只是不再低头。镜头缓缓后拉,她独自立在空旷天台正中,铁丝网外阴云翻涌,画面定格。【环境音效与台词】全程无背景音乐,原生天台空旷混响,风声持续作底: · 开场压抑:风穿过铁丝网的呜呜声,人群逼近的杂乱脚步与嗤笑低语(内容模糊不清),书包砸地的一声闷响,课本散落的哗啦声 · 爆发瞬间:她蹬地欺身的一声鞋底摩擦,第一拳命中的厚重闷响瞬间盖过所有嘈杂 · 六连切:拳肘命中的闷响错落密集,扫踢带起的低沉风声,摔地与滑地的连续钝响;两次微慢段所有声音骤然压低拉长,只剩衣料绷紧声、断发飘落的极轻声与她一次沉稳的呼吸,弹回极速时音效同步炸回 · 群体溃退:杂乱仓皇的脚步与爬地声,天台铁门撞开又弹回的哐当声,随后只剩风声 · 收尾:课本拾起的纸页摩擦声,拍灰的两下轻响,理领结的细微布料声,风声渐大,收在一声干净的定格静默 · 人物状态:她全程不发一言,仅出招瞬间伴随极短促的吐气声;终镜前一次绵长的呼气,平静收场
How to Use This Prompt
Copy the prompt
Grab the cleaned-up prompt in your language, or send it straight to the editor with one click.
Pick your model
We pre-select the model that produced this result, but you can swap it for any model Vorla supports.
Generate and refine
Run it as-is to reproduce the look, then edit the wording to make the result your own.
More from Seedance 2.0
Cinematic Framework: Shot in the style of a 1950s major-studio romantic drama. 4:3 Academy ratio. Rich black-and-white 35mm cinematography on fine-grain stock, high contrast with silvery highlights and velvety blacks. Classical Hollywood composition — balanced framing, generous headroom, actors staged in defined planes. Camera mounted on smooth dolly tracks; all movement is deliberate and gliding. Deep-focus wide shots in the master, creamy shallow focus in close-ups with circular bokeh. Slight diffusion on faces for that period studio glow. High-key lighting with soft, motivated shadows. No handheld camera work. No modern coverage. Setting: A busy American city street, early 1930s. Cobblestones, red-brick storefronts with hand-painted signage, a parked Model A Ford. Pedestrians in wool overcoats, cloche hats, and flat caps move with purpose in both directions, naturally avoiding the central action without looking at the lens. Mid-morning overcast, practical daylight feel. 0:00–0:03 — The Interception (Wide Master)Wide shot, eye level, deep focus. The woman, late twenties, dark wool day coat and small hat, walks toward camera through the crowd, both hands gripping a woven basket of fresh apples and pears. A young man in his late twenties, three-piece suit, tie, and wool overcoat, enters from frame left and steps calmly into her path. She slows. He stops, removes his hat with his right hand, and executes a single, clean bow from the waist — formal, unhurried. Pedestrians flow around them. Hold the wide shot for the full beat; let the gesture play out in real space. 0:03–0:08 — The Plea (Reverse Close-Up)Cut to medium close-up, classical over-the-shoulder reverse angle, shallow focus. The man looks up from his bow, his face now eye-level with hers. Sustained eye contact, earnest and slightly nervous. He speaks, warm and clear, natural mouth movement: "My lady, may I walk with you and carry your basket?" A slight swallow after the final word. The camera executes a slow, imperceptible dolly push-in — the 1950s dramatic emphasis — closing the distance by mere inches as he finishes the line. 0:08–0:11 — The Decision (Reaction Close-Up)Reverse shot, close-up on the woman. Same shallow focus; the street behind her dissolves into soft, moving bokeh. She holds his gaze for a full beat, assessing him. Her expression shifts from surprise to curiosity to genuine warmth. A small smile breaks, reaching her eyes. She tilts her head slightly — restrained, composed, emotionally true to the studio tradition. No dialogue. Let the smile do the work. 0:11–0:15 — The Handoff (Tracking Two-Shot)Cut to medium two-shot, balanced 4:3 composition, both actors centered with classical headroom. She extends the basket toward him with both hands. He receives it, their fingers grazing. They turn together in synchronized movement — deliberate studio blocking — and begin walking up the sidewalk. The camera trucks smoothly alongside them at an exact walking pace, holding the two-shot as they move deeper into the street traffic. The crowd closes behind them. Hold on the moving shot through the final frame; do not cut to black. Performance & Texture: Naturalistic acting in the 1950s studio register — emotional sincerity without theatrical exaggeration. Realistic skin texture, visible fabric weight in wool and tweed, woven basket detail. Continuous ambient sound: cobblestone footsteps, distant car horn, indistinct crowd murmur, fabric rustle. No music. No film-grain overlay beyond the native 35mm texture. No modern editorial tricks.
Create a fifteen second ultra realistic social beauty vlog in a cinematic 16:9 frame, four K detail, thirty frames per second, natural real time motion. Feature an original young adult woman with long blonde hair arranged in a neat low ponytail, natural luminous makeup, detailed blue eyes, pale pink satin pajama shirt, layered silver necklaces, rings, dark burgundy manicure, and subtle small hand tattoos. She sits close to an eye level camera in a bright modern living room. A tall snow dusted Christmas tree decorated with soft warm white lights fills the right side of the background. Large windows provide gentle cool daylight while warm reflected room light softly shapes her face. Use a fixed camera on a small tabletop tripod with the natural perspective of a modern phone camera, approximately thirty five millimeter equivalent. Frame her from the upper chest upward, positioned slightly left of center so the Christmas tree balances the composition. Preserve realistic skin pores, tiny facial hairs, individual eyelashes, fine flyaway hairs, accurate jewelry reflections, natural fabric folds, subtle breathing, irregular blinking, believable finger pressure, and physically correct object weight. Include mild autofocus breathing, small automatic exposure adjustments, authentic motion blur, restrained digital sharpening, and very subtle camera vibration. During the first three seconds she looks into the lens with a relaxed welcoming smile, gently tilts her head, adjusts one loose strand of hair, and glances toward a product below the frame. From seconds three to six she reaches downward with both hands and lifts a large pastel pink circular beauty advent case into view. She holds it beside her face and slowly rotates it toward the camera as glossy details catch the window light. From seconds six to nine she opens one small compartment with careful finger movement and reveals a pale pink cleanser bottle. Her expression changes into genuine delighted surprise. The focus naturally shifts from her eyes to the bottle as she brings it closer to the lens, then smoothly returns to her face when she pulls it back. From seconds nine to twelve she lowers the bottle and opens a coordinated makeup palette containing neutral eyeshadows, a soft rose blush, and a champagne highlighter. She angles the palette so each powder surface reflects a slightly different amount of light. Maintain complete product continuity and accurate hinge movement. From seconds twelve to fifteen she holds the open palette beside her cheek, gives the camera a warm knowing smile, then gently closes the case as the camera makes an almost invisible slow push forward. Finish on her eyes, the pink products, and softly glowing Christmas lights. Add an elegant upbeat beauty pop instrumental at approximately one hundred and four beats per minute with airy synthesizer, soft bells, muted hand claps, rounded bass, delicate percussion, and no vocals. Synchronize gentle musical accents with the case reveal, cleanser reveal, and final palette presentation. Include quiet room ambience, satin movement, compartment clicks, packaging sounds, and soft fingernail taps beneath the music. No spoken dialogue, subtitles, captions, logos, watermarks, artificial beauty filter, plastic skin, excessive background blur, face distortion, changing identity, extra fingers, fused hands, warped jewelry, duplicated products, floating objects, inconsistent packaging, hair flicker, background movement, abrupt exposure changes, jump cuts, or unnatural camera motion.
Scene start: shows the same shot as the reference image. The translucent snake moves along the girl's arm, displaying a faint, shifting bioluminescence. Scene narrative: The woman relax says: "There is an archon above your head. Get rid of it." Scene end: The snake suddenly bites the camera with an agile, predatory movement. Cinematic, photorealistic scene.
Use @[storyboard ref] as a strict P01-P09 path: each panel's angle, framing, pose, and contact geometry is a required fleeting waypoint, reached once in order without skips or merges. Start on exact P01; one continuous track passes through P02-P09 without holds or cuts, ignoring sketch artifacts while written fields and @[character ref] define final render. Use @[character ref] only for Suzune's final identity, wardrobe, right gauntlet, and katana. C2, C5, and C6 are separate six-limbed cobalt Razorclaws with pale bellies and long foreclaws. C3, C4, and C7 are separate slim cobalt Razorwings with turquoise wings, forked violet tails, and yellow eyes. Use one 20mm virtual lens throughout the uninterrupted shot. LOOK + WORLD: Hand-drawn cel-shaded 2D anime, dry-brush ink; storm void, shrine bridge, torii, four falling slabs; coral-cobalt light, 20mm deep focus, dust. ACTION + CAMERA: Pro Anime Combat: elastic compressions, instant rebounds, glancing cuts, continuous carry-through. Camera never settles; one attacking or evading target per cut, no holds, hit-stops, slow motion, freezes, resets, or teleports. VFX SYSTEM: A coral edge wake stays on the katana; each cut releases one cobalt-fluid streak and cyan fragments, then clears with no ranged wave. SOUND: Sparse breakbeat pulse; storm wind, rock fractures, wing snaps, blade impacts, creature cries. MOTION PHRASES: P01-P09: Ground front 3/4 FS, continuous developing track with one forward handoff: P01 is the literal first frame, Suzune's bare left foot filling the foreground on the first slab while C2 lunges from upper right; the camera instantly flows into P02's low profile as she rebounds beneath C2's airborne claw swipe and passes one horizontal cut through it, hair and skirt trailing without reset; without settling, it wraps into P03's low rear 3/4 as her right foot brushes the second slab and C3 dives from upper right, then curves beneath her into P04's ground front 3/4 while the same rebound rises through C3 with one uninterrupted cut; the path climbs to P05's high side view as her left toes graze the third slab and her waist rotation crosses C4's attacking S-turn with one crosscut; it drops behind into P06's rear 3/4 as her right foot touches the fourth rock's vertical face and wall-running C5 catches the katana between moving foreclaws for only the sampled waypoint, then glides into P07's tight profile while Suzune rolls through that contact and reverse-cuts C5 without stopping; the camera slides under the slab into P08's low side view as inverted C6 lunges and her continuing body arc slices through it once; one forward handoff rises into P09's high front 3/4 as C7 corkscrews beak-first, Suzune passes one cut through it and keeps falling down-right toward more moving creatures; reach every named panel composition once, never interpolate an alternate pose, and let each source-bound coral edge wake, cobalt-fluid streak, and cyan fragment burst clear during the ongoing motion.
Seedance 2 prompt. Create a 15-second cinematic breakup scene in strict 16:9 landscape, 1080p, realistic Korean-drama emotional close-up style. Both characters are adult Asian man and woman in their twenties. REFERENCE: image1 = main adult woman reference. Use image1 as the strict reference for the woman’s face, facial proportions, hairstyle, hair color, outfit, color palette, age impression, body type, skin tone, styling, visible accessories, and overall recognizable vibe. Do not invent a new outfit, hairstyle, hair color, makeup style, accessory, or body type. Preserve all visible clothing, hair, accessories, and styling from image1 consistently. Her face must stay sharp, clear, and recognizable in every shot. SCENE: A quiet riverside pedestrian path or small neighborhood bridge at night. Soft reflections on dark water, distant city lights, blurred green trees, and warm streetlamp bokeh. Lonely, calm, emotionally heavy atmosphere, like a final goodbye in a quiet city corner. Use an 85mm portrait telephoto lens, very shallow depth of field, warm soft night lighting, subtle film grain, and premium Korean TV drama realism. Camera stays stable with only extremely subtle breathing-like movement. CHARACTERS: The woman from image1 is the absolute emotional focus. Her face must remain in focus at all times, with realistic skin texture and natural facial detail. Do not over-retouch her skin. The adult man stands on the right side as a blurred foreground figure, always facing away from the camera. Only his blurred back of head, shoulder, and one hand are visible. Do not show his clear face. He must never steal focus. STORY: This is their final meeting after deciding to break up. The woman is trying to stay calm, polite, and dignified, but she is close to breaking. She still loves him, but knows this goodbye is final. The man remains silent. His only action is one slow, gentle touch on her cheek. The emotional core is quiet heartbreak: she tries to smile so the goodbye will hurt him less, but every small expression shows she is falling apart inside. EMOTIONAL ACTING RULE: Do not make the scene empty, still, or expressionless. The woman must show continuous micro-emotional changes throughout the full 15 seconds. Her face must never stay frozen. Use layered emotional acting: wet lower eyelids, slight eye tremble, tiny failed smile, subtle lip quiver, jaw tension, small swallow, breath catching, brief downward glances, a smile that appears and collapses, one quiet almost-cry moment, then a fragile smile returning. The sadness must feel deeper than simple crying. She is trying not to cry, but her face keeps betraying her. TIMELINE: 0:00–0:02.5 Tight close-up on the woman’s face. The man’s blurred shoulder and back of head remain on the right foreground. She looks at him quietly with wet eyes. Her expression is calm on the surface, but wounded and exhausted underneath. Her lips part slightly as if she wants to say something, but no words come out. A tiny breath gets caught in her throat. She blinks once, slowly, trying to hold back tears. Her lower lip trembles very slightly, then she presses it still. For a brief moment, her eyes look like she is silently asking him not to leave. 0:02.5–0:05.0 She lowers her gaze. Her eyelashes look slightly wet. She presses her lips together and lightly tightens her jaw, trying to regain control. Her throat moves with a small swallow. She inhales softly, but the breath is uneven. She looks back up and tries to smile. The first smile fails halfway. The corners of her mouth lift for a second, then fall again because the sadness is too heavy. Her eyes remain glossy and painful, full of unsaid words. 0:05.0–0:07.5 She tries again to smile, weaker this time. The smile is faint, fragile, almost apologetic. Her eyes do not match the smile; they look hurt, regretful, and full of things she cannot say. She gives a tiny breathy exhale, almost like a quiet laugh that breaks before becoming a laugh. Her expression should say: “I wanted us to be okay, but I know we are not.†She holds eye contact, then briefly looks down because looking at him hurts too much. 0:07.5–0:10.0 The man remains silent. From the blurred right side, he slowly raises one hand and gently places his palm on her left cheek. The movement must be slow, natural, and continuous. His fingers must look anatomically correct and must not cover her eyes. When his palm touches her cheek, her expression visibly cracks. Her eyelids lower, her breath stops briefly, and her cheek instinctively leans slightly into his palm before she realizes it. This unconscious lean should feel heartbreaking, as if her body still wants to stay even though she has decided to let go. Her eyes become shinier, but no dramatic tear stream. She almost cries, then catches herself. 0:10.0–0:12.5 She looks back at him with tears in her eyes. A painful smile slowly appears. It starts restrained and helpless, then becomes a delicate cry-smile. Her eyebrows lift slightly in sadness. Her mouth tries to smile, but the smile trembles. The smile should feel like she is thanking him, forgiving him, and saying goodbye at the same time. The man’s hand stays on her cheek briefly, then slowly lowers out of frame. When his hand leaves, her expression almost collapses. Her smile fades, her eyes drop for half a second, and she looks emptier, as if the final warmth has disappeared. Then she forces herself to hold on. 0:12.5–0:15.0 The camera pulls back very slightly to a medium close-up above the chest, still keeping the woman as the absolute focus. She takes a small shaky breath, looks at him with wet eyes, and says softly in Korean: “앞으로 잘 지내.†Her voice should be gentle but slightly broken, like she is saying it with the last bit of strength she has. After saying it, her lips stay slightly parted for a beat. Her smile fades for half a second, revealing the real pain underneath. A tiny tear may gather at the lower eyelid, but it should not pour down dramatically. Then she forces one last fragile sad smile. She briefly looks down, gathers herself, then looks back at him. End naturally on her tearful smile, with heartbreak, affection, regret, and reluctant acceptance mixed together. Do not cut to black. Do not add hugging, kissing, walking away, or extra dialogue. PERFORMANCE REQUIREMENTS: The emotional change must be detailed, slow, and continuous. Her eyes and mouth must act differently: her eyes reveal the pain first, while her mouth tries to hide it with a smile. Every smile should look like it costs her effort. The sadness should feel restrained but deeply visible. No sudden emotional explosion, loud sobbing, screaming, dramatic shaking, instant face change, blank stare, or frozen expression. She must look like a real person desperately trying to stay composed during a final goodbye. FACIAL DETAIL: Show realistic micro-expressions: wet lower eyelids, tiny mouth tension, subtle lower-lip tremble, small swallow, breath catching, forced smile that fails, brief almost-cry moment when his hand touches her cheek, hollow pause after his hand leaves, brief emptiness after the final line, then a weak sad smile returning. Keep expressions natural, not theatrical. SADNESS INTENSITY: Make the sadness emotionally clear. The audience should feel she is devastated but trying to be kind and composed. Her sadness must come from restraint, not loud crying. Use trembling eyes, failed smiles, uneven breathing, and small pauses. The final expression must feel like someone smiling only because crying would make the goodbye harder. CAMERA RULES: This is an emotional acting scene, not an action scene. Use mostly stable close-up framing. Allow only extremely subtle breathing-like camera movement and one very slight pull-back near the end. No fast push-in, orbiting camera, dramatic handheld shake, frequent cuts, flashy cinematography, or rapid zoom. The camera should feel intimate, quiet, and respectful. VISUAL STYLE: Live-action Korean drama realism. Warm night lighting. Very shallow depth of field. 85mm telephoto portrait compression. The woman’s face stays sharp and clear. The man remains blurred the entire time. Natural warm bokeh lights in the background with faint riverside or bridge reflections. Soft highlights on her wet eyes. Realistic skin texture. No plastic beauty filter, no over-smoothed skin, no glossy commercial look. AUDIO: No background music. Use only natural scene audio: soft night ambience, faint distant city hum, subtle cloth movement, gentle breathing, small pauses, quiet breath catching, slight shaky inhale, soft sound of his hand touching her cheek, and one soft Korean line near the end: “앞으로 잘 지내.†NEGATIVE: No subtitles, watermark, short-video beauty filter, plastic skin, over-smoothed skin, anime look, cartoon look, advertisement look, studio glamour pose. No sudden heavy crying, screaming, exaggerated tears, tears pouring like water, snot, face twitching, frequent blinking, blank staring, instant expression switching, sudden laughing, abnormal mouth shapes, or deformed teeth. The man must not turn around or reveal his face, steal focus, hug her, kiss her, move closer suddenly, wipe her tears, or pinch her face. He may only perform one slow gentle cheek touch. No extra fingers, fused fingers, broken fingers, hand passing through the face, drifting touch position, palm covering her eyes, or sudden slap-like movement. No camera orbit, violent handheld shake, frequent cuts, flickering background, character position drift, hairstyle change, outfit change, identity drift, face change, clothing change, or hair color change.
15-second, 16:9 premium technology launch film for an ordinary carrot. Pure black background, immaculate studio lighting, cinematic macro photography, ultra-slow rotation, minimalist white sans-serif titles, ethereal electronic score. Treat the absurd subject with total seriousness and restraint. Keep the same photorealistic carrot throughout: rich orange-red skin, fine vertical ridges, short fresh green stems. [0–4s] A full, straight carrot floats and slowly rotates in darkness. Cool white rim light defines its silhouette; a soft warm highlight moves across the skin. Macro droplets roll slowly over the surface. Fade in: “Carrot Pro” “Remarkably fresh.” [4–8s] Three fast studio macro shots: 1. A moist cross-section reveals radial fibers and subtle rings under precise top light. Title: “Vibrant.” 2. The long curve and fine ridges form a sculptural silhouette in side backlight. Title: “Crisp.” 3. Fine mist lands on the skin; droplets cling to the grooves in ultra-slow motion. Title: “Sweet.” [8–12s] The carrot turns gracefully in midair; its green stems respond with natural inertia. A clean human hand enters from below, catches it securely, then rotates it with believable weight and grip. A restrained orange-gold light trail traces the silhouette as the music reaches a monumental climax. Hero shot like revolutionary flagship hardware. Display: “The most powerful carrot ever.” “20% crunchier than the previous generation.” [12–15s] The music resolves. The same carrot stands upright, floating alone at center frame with generous negative space. Display: “From $3” “Carrot Pro. Rooted in flavor.” End with one clean electronic chime and hold the final frame like a product poster. No dialogue. Accurate specified text only. No extra subtitles, logos, kitchen, utensils, soil, packaging, anthropomorphism, or deformed hands. Preserve carrot shape, color, and proportions across every shot.
All Video Categories
Grok Imagine
Pink and Cyan Liquid Text Intro
Find Your Next AI Prompt
One search covers every model and both mediums, and copying what you find costs nothing.

















