AI Video Prompts for First-Person & POV

SUBJECTS Female enforcer: White long hair, with slightly fluorescent-colored tips; Wearing a loose jacket, fitted top, dark tight pants, and glowing speed shoes emitting a cyan light; Target male: Male, wearing ordinary clothing, a thief; ENVIRONMENT Night market setting, stalls arranged irregularly, string lights at uneven heights, crowd moving in inconsistent directions, wet reflective ground, distant city skyscrapers with neon lights Timeline SHOT 1 (0:00-0:02) Third-person wide shot, 35mm, slight handheld The female protagonist jumps from the edge of a rooftop and enters the frame, body leaning forward as she falls, the city compressed below; the camera follows slightly delayed, creating a sense of speed difference SHOT 2 (0:02-0:03) Medium follow shot During the fall, force is applied through the feet, the shoe soles light up instantly, the fall transitions into a forward dash, the body is pushed and leveled out, the camera passively catches up SHOT 3 (0:03-0:05) POV, 18mm Hands enter the frame, rapidly moving between skyscrapers, buildings stretch into irregular color blocks, space appears like brushstroke smears SHOT 4 (0:05-0:08) POV continues Descending above traffic flow, weaving through gaps between vehicles, headlights form flowing light trails, slight camera shake, focus lags slightly SHOT 5 (0:08-0:10) Third-person rear-side view The protagonist slides into the night market, shoe glow gradually fades, speed decreases but does not fully stop, merging into the flow of the crowd, the camera gets interrupted by passersby blocking the view SHOT 6 (0:10-0:12) Handheld medium shot, obstructed view The camera follows the protagonist moving, the crowd continuously creates occlusion; the target male appears through gaps in the crowd: approaching someone’s backpack, hand reaches out to test then pulls back, head lowered pretending nothing is happening while eyes glance upward to observe; takes two steps, stops again, quickly scans surroundings, behavior clearly suspicious SHOT 7 (0:12-0:13) Rapid push-in The protagonist instantly lowers her body and accelerates, bursting straight out of the crowd to lock onto the target, movement without warning SHOT 8 (0:13-0:14) Medium shot Aerial flying kick hits the male, knocking him to the ground, movement clean and decisive, no unnecessary pause SHOT 9 (0:14-0:15) Target POV looking upward A hand enters from above, naturally lowering a police badge into view; the camera auto-focus locks onto the metallic details of the badge, the night market background fully blurs into multicolored bokeh, frame holds still

FPV sprint through a dense city as buildings begin collapsing, camera racing at ground level between cars and debris, weaving through falling concrete chunks, ducking under collapsing scaffolding, then accelerating into a narrow alley as a shockwave explodes behind, fire and dust chasing the camera, flipping briefly as the ground cracks open, then stabilizing and bursting out into a wide street engulfed in chaos, ultra-fast, relentless motion. Epic, cinematic

FORMAT: 15 seconds / 145 BPM / 15 beat-synced shots SUBJECT: @[image1] ENVIRONMENT: Apartment → bathroom → kitchen → taxi → airport terminal → security check → boarding gate → airplane cabin → hotel room night MOOD ARC: Late wake-up → high-pressure travel rush → relief → quiet exhaustion SHOT CHANGES (key differences): • Shot 11: exits apartment, jumps into fast-moving taxi (rain streaks on window) • Shot 12: runs through airport terminal with suitcase, departure board flashing • Shot 13: boarding pass scanned at gate • Shot 14: airplane window seat, city lights below • Shot 15: hotel bed collapse, suitcase still half-open

Create a photorealistic ultra-cinematic 15-second first-person POV action scene. We are riding on the back of a massive black dragon racing through a storm-torn canyon at night. Both hands remain visible in frame, gripping the dragon’s wet armored neck ridges like living reins. Wind lashes hard, rain cuts sideways, lightning reveals jagged cliffs and abyss drops below. The tone is epic, intense, and premium blockbuster fantasy realism. 0-3s: first-person POV low over the dragon’s neck, both hands gripping tightly as the dragon surges through the canyon at high speed, black scales slick with rain, breath vapor and embers trailing from its mouth, cliffs rushing by on both sides. 3-6s: ahead in the storm, a squad of armored aerial hunters appears and attacks, launching multiple flaming bolt-spears directly toward us. The projectiles race through the rain toward camera while the dragon banks violently between cliff walls. 6-9s: at the last possible instant, one hand releases the dragon and rises into frame with a sharp command gesture. Time stops instantly. All incoming bolt-spears freeze in mid-air just in front of the POV, along with rain droplets, sparks, flames, and drifting mist, suspended in a rippling invisible field. 9-12s: the raised hand slowly twists sideways. The frozen incoming projectiles rotate in the air one by one until all of them are now pointing back toward the attacking riders. The enemies remain moving in the distance, but the captured weapons are fully controlled and aimed at them. 12-15s: with a second precise hand gesture, time snaps forward only for the redirected projectiles, sending the entire suspended volley back at extreme speed into the attacking aerial hunters. The returning flaming bolt-spears slam into the enemy formation in the storm as the dragon drives forward through the chaos. End on a powerful forward-moving POV frame with one hand returning to grip the dragon.

extremely fast-paced cinematic FPV flying through the ancient Panchala kingdom in India, hyper-realistic, 4K detail, golden sunrise lighting, vast royal palaces, intricate sandstone architecture, bustling markets, warriors training, horses galloping, temples, dense forests and rivers surrounding the city, dramatic camera dives, sharp turns through narrow corridors, epic scale, volumetric lighting, rich warm color grading, historical authenticity, highly detailed textures, immersive, film-grade realism

15-second continuous single-shot action sequence. No cuts. No scene transitions. First-person POV cinematic realism. Color palette: cold overcast sky, mud browns, steel greys, desaturated colors, dark blood tones. Scene: Open battlefield under cloudy sky. Mud, churned ground, scattered infantry formations, distant banners barely visible through dust and smoke. 0–3s — approach POV from mounted knight. Horse breathing heavy beneath. Subtle up-and-down motion. Hands gripping reins and lance. Other cavalry riders flanking left and right. Distant enemy infantry line ahead. 3–6s — charge begins Horse surges forward. Camera shakes rhythmically with gallop. Mud splashes upward. Nearby riders shouting (muffled). Formation tightening as speed increases. 6–10s — full speed Ground rushing past. Lance steady but vibrating slightly. Enemy line growing fast — shields, spears becoming distinct. Arrows begin falling randomly. One rider nearby hit, falling out of frame. 10–13s — impact Violent collision with infantry line. Shield hits lance — splintering wood. Camera jolts hard sideways. Horse pushes through bodies. Brief glimpses of faces, metal, chaos at arm’s length. 13–15s — aftermath Movement becomes unstable. Breathing heavy. Horse slows slightly but still moving forward through broken formation. Smoke and dust filling view. Vision partially obscured. Freeze. Style: Raw, grounded, chaotic realism. Limited visibility. Physical impact. No clean choreography. Avoid: cinematic perfection, wide clean shots, fantasy armor, exaggerated effects.

SUBJECT: A tired woman in @ image1 in a loose tank top and sleep shorts, slow habitual movement. Slightly smeared eyeliner, bare feet, heavy posture, detached face. ENVIRONMENT: A cramped cluttered apartment with an unmade mattress, scattered clothes, a narrow hallway, a damp bathroom with dim tile reflections, and a tiny kitchen crowded with dirty dishes and empty bottles. Warm practical lamps mix with sickly green neon leaking through blinds and door glass, turning the rooms into a humid late-night maze. MOOD: Detached routine turns quietly uncanny, as if an unseen presence is floating above her and waiting for her to notice. COLOR LOGIC: Matrix Green Look CAMERA: POV overhead follow in a strict bird's eye view, locked directly above the top of her head at all times, perfectly centered over her body from start to finish, floating smoothly with no shake, tilt, angle drift, or side offset, passing through ceilings and door frames as one uninterrupted camera event. 24mm wide, digital clean look, locked overhead tracking package throughout. SCENE: She wakes on the mattress. Sits up under the lens. Still centered under the lens, reaches to the floor. Picks up the cigarette and lighter. Places the cigarette between her lips. Lights it. Drops the lighter on the mattress. Stands up with the cigarette still with her. Under the same overhead lock, crosses the cluttered room. The lens tracks directly above her into the narrow hallway. Enters the bathroom. Still pinned overhead, keeps the cigarette in her right hand. Extends that arm away from the running water. Leans on the sink. Turns on the tap with her left hand. Splashes water onto her face with her left hand. The overhead follow carries her back out of the bathroom into the hallway. Moves along the hallway past the bathroom door. Turns into the tiny kitchen. Still centered under the camera, keeps the cigarette with her. Reaches across the dirty dishes with her free hand. Picks up a glass from the counter. Stops exactly under the lens. Holds the cigarette in her right hand and the glass in her left hand. The lens stays fixed above her. Freezes. Looks right. Looks left. Takes a drag from the cigarette. Snaps her head straight up into the lens. Blows smoke toward the camera. Locks eye contact. SFX: lighter flick, inhale, faint city hum, refrigerator buzz, soft bare footsteps, water run. Sodium amber particles, toxic green neon reflect off tile, smoke, bottles, and damp surface.

High-speed, continuous FPV drone shot navigating through the vibrant, authentic backstreets of Istanbul. The sequence begins right beside a traditional street-side meatball (köfte) vendor, with thick smoke rising from the grill. The camera seamlessly and rapidly weaves through a group of energetic children playing street football in a narrow, cobbled neighborhood alley. Without losing any momentum, the FPV drone dives through the open front door of a bustling local cafe, gliding smoothly just inches above a wooden table filled with authentic Turkish dishes, tea glasses, and chatting patrons. It fluidly exits the cafe by flying through an open window, zips past a live street photography session with a photographer taking portraits, and then rapidly ascends straight up into the sky to reveal a breathtaking, sweeping panoramic view of the Istanbul cityscape and the Bosphorus. Cinematic photorealism, 8k resolution, highly detailed, dynamic natural daylight, warm color palette, motion blur conveying extreme speed, flawless transitions, no camera stutter. Accompanied by the native audio of sizzling meat on a grill, children laughing and kicking a ball, ambient cafe chatter, clinking tea glasses, rapid camera shutter clicks, and the intense rushing wind of the drone flight transitioning into a sweeping, majestic cinematic orchestral swell. Negative: cut, slow motion, blurry, low resolution, bad physics, impossible camera collisions.

Use @[image1] as Veil'sreference. Use @[image2] as Jing's reference. The characters must strictly match the character sheet. Use @[image3] as storyboard reference. Treat each storyboard panel as a sequential keyframe. Preserve the shot order and make the video feel like the storyboard has been translated into continuous live-action motion. Create a 15s POV martial-arts fight between Veil and Jing. Sequence: 1. Jing POV: Veil launches a huge distorted high kick directly into lens. 2. Veil POV: Jing rushes camera with a brutal punch. 3. Jing POV: Foreground block and fast counter strike toward camera space. 4. Veil POV: Jing grabs near camera during chaotic close-range struggle. 5. Jing POV: Rapid punches invade camera space from Veil. 6. Veil POV: Jing lands elbow strike and violent palm push. Camera snaps backward. 7. Jing POV: Veil attacks aggressively with near-lens kicks and punches. 8. Veil POV: Final chaotic exchange. Hands, cloth, feet and motion blur constantly invade foreground. Aggressive first-person combat filmed with violent handheld camera movement and extreme ultra-wide lens distortion. The camera constantly switches between each fighter’s eyes. Only one fighter is fully visible at a time. The opponent appears only as foreground hands, blurred limbs, partial blocks and near-lens motion streaks. Pure white background. Black cinematic bars top and bottom. Strong 3D frame-break effect. Hands extending above the top bar, feet crossing below the bottom bar, hair, cloth and motion blur spilling outside frame toward viewer. High-end stylized 3D realism, dynamic cloth simulation, heavy motion blur, sweat, dust, debris particles, impact shake, intense cinematic lighting, claustrophobic framing, chaotic close-range combat.

First-person action-camera POV, wide-angle lens with mild distortion — NOT a fisheye: no black or vignetted corners — camera mounted at the wearer's chest: very shaky handheld throughout — constant hard jitter and bounce, the frame bobbing and swaying violently with every step, whipping hard on impacts and sprints, never smooth, never stabilized, never tripod-still. The wearer is a medieval peasant man; his weathered hands with dirt-caked nails and coarse brown wool sleeves enter the frame constantly and stay consistent in every shot. Cold medieval England, early-morning soft light, wet deep mud everywhere with sheen and ruts, chimney smoke, thatched and timber-framed village below a massive grey stone castle on a hill. Mud behaves physically: splatter arcs, heavy clinging clumps, wet squelch. Raw unedited camera footage, natural light only, no color grading, not cinematic. No on-screen text, no subtitles. No background music, diegetic sound only. No cuts within a shot. the woman is @ Shot: First-person action-camera POV, wide-angle lens with mild distortion (no fisheye, no black corners). Waking at dawn in a timber barn loft, thin blades of light through gaps in the plank walls, straw and grey fur. The wearer's head turns and his young wife is right there beside and close him — early 20s, loose brown braid, freckled skin — just waking herself: she rubs one eye with the back of her hand, breaks into a wide sleepy yawn, stretches a little, blinks slowly and settles into a soft drowsy smile. She murmurs, in a weak, soft, just-woken sleepy voice, half a yawn still in it, looking into the lens (on screen, lip-synced): "Stay a while... the day can wait." From behind the camera he answers, warm and low, off-screen POV (no lip-sync): "The lord's fields can't, love." She gives a lazy grin and hooks a hand behind his neck to pull him closer, his weathered hand rising into frame to her cheek, and she teases, still sleepy and soft: "Then let the lord plough his own field." He leans in and kisses her, the frame dipping in close, her hand sliding to the back of his head, a soft shared laugh. Warm, unhurried, natural soft daylight. No cuts within a shot. No background music. First-person action-camera POV, wide-angle lens with mild distortion — NOT a fisheye: no black or vignetted corners — camera mounted at the wearer's chest: very shaky handheld throughout, constant hard jitter and bounce, the frame bobbing and swaying violently with every step, whipping hard on impacts and sprints, never smooth, never stabilized. The wearer is a medieval peasant man; his weathered hands with dirt-caked nails and coarse brown wool sleeves enter the frame constantly and stay consistent. Medieval England, cold damp climate, wet deep mud with sheen and ruts, chimney smoke, thatched and timber-framed village below a massive grey stone castle on a hill; natural light only. Mud behaves physically: splatter arcs, heavy clinging clumps, wet squelch. Raw unedited camera footage, no color grading, not cinematic. No subtitles, no on-screen text, no background music (diegetic sound only). No cuts within a shot. Shot 1: Cold early morning. Looking down at a deep-rutted mud road while walking, boots squelching, wet mud glistening; the wearer's weathered right hand holds a gnarled walking stick and a small tied burlap sack. The camera tilts up mid-stride revealing a massive grey stone castle with flags on a hill above a medieval village, horse carts and hooded villagers on the road ahead; a passing villager brushes close past the lens. Frame bobs with every step. Shot 2: Same POV arrives at a muddy market square. A bald man with a thick reddish-grey beard in knee-length chainmail and a grey cloak stands holding a tall staff beside a wooden cart piled with bulging burlap grain sacks, a tray of eggs and live brown chickens; a clerk's hand writes on parchment at the right edge of frame; crowd murmurs. The wearer's weathered hands reach into frame and cinch a burlap sack shut on the cart edge. Shot 3: Same POV. An old woman in a dark grey hooded shawl, 70s, deeply lined face, bushy grey brows, shrewd eager eyes, hurries up close through the mud; she points a bony finger down at the eggs and looks up at the lens, asking in a cracked voice (on screen, lip-synced): "How much for five?" From behind the camera the wearer answers plainly, off-screen POV (no lip-sync, face not seen): "A penny." She digs a small coin from a cloth purse and presses it into the wearer's open weathered palm as his hand comes into frame to receive it. Behind her hang furs on a stall, clay pots, a pig rooting in the mud.

【风格】真实手机自拍生活流（Handheld Selfie POV, iPhone Front Camera Look），非影视打光，真实摄影（Photorealistic），4K高清，轻微手机视频噪点（subtle video grain），皮肤质感真实无磨皮 【时长】10秒 【场景】温馨卧室床上，粉色绒面床头板，豹纹图案床单，红色卡通玩偶枕头，周围散落米色毛绒玩具，暖黄色室内灯光从右侧打来 【角色】她@图片1（侧躺在床上抱着枕头的女性，深色头发松散扎起、碎发垂落，穿米色短袖上衣） [00:00-00:03] 镜头1：怼脸对视（Extreme Close-up, Handheld POV） 男友第一人称手持自拍视角，镜头怼脸距离极近，画面带轻微手臂呼吸感晃动。 她@图片1 侧躺着看向镜头，眼神直视，睫毛缓慢眨动，嘴角慢慢咧开笑意。 特写细节：脸颊自然泛红，鼻尖有暖光高光，皮肤毛孔和绒毛在近距离下清晰可见，几缕碎发贴在额头和脸颊上。 [00:03-00:06] 镜头2：娇羞埋头（Shy Reaction） 她突然憋不住笑，抬起一只手捂住头顶，整张脸埋进旁边的红色玩偶枕头里。 肩膀因为闷笑一耸一耸，手指轻轻抓住枕头边缘，枕头绒面被压出真实褶皱。 镜头跟着她的动作小幅下移调整角度，手持晃动加大一点，仿佛拍摄者笑着凑近。 音效：被枕头闷住的轻笑声，布料摩擦声。 [00:06-00:10] 镜头3：回头偷看（Turn Back, Hold on Face） 她把脸从枕头里慢慢转出来，先露出一只眼睛偷看镜头，然后整张脸转过来。 特写：眼眶带着笑出的湿润光泽，脸颊红晕加深，头发被枕头蹭乱了几缕翘起。 她抿嘴一笑，又轻轻眨了一下眼，镜头保持怼脸距离定格在她脸上收尾。 音效：轻声的鼻音笑，床单窸窣声。

Seedance 2.0 Prompt (15s Version) Format: Continuous single-take police bodycam Audio: No music, only ambient + AI-generated dialogue Style: Hyper-realistic, raw, live-action bodycam 0:00 – 0:03: Bodycam starts abruptly. Officer stands outside a dark, abandoned house at night. Flashlight beam scans broken windows and overgrown yard. Subtle breathing, radio static. Officer quietly speaks (AI-generated dialogue). 0:03 – 0:06: Door creaks open. Officer steps inside. Flashlight reveals dust, decay, and empty rooms. Floorboards creak. Officer calls out cautiously (AI-generated dialogue). 0:06 – 0:09: Quick, tense sweep of interior. Camera movement slightly shaky. Strange distant noise (ambiguous source). Breathing intensifies. Officer mutters something uneasy (AI-generated dialogue). 0:09 – 0:12: Officer reaches staircase and ascends. Each step creaks loudly. Flashlight points upward into darkness. At the top, he turns into a pitch-black hallway. 0:12 – 0:14: As he turns, a tall, unnatural, shadow-black humanoid figure briefly appears at the edge of the flashlight beam—subtle, distorted, barely visible but deeply unsettling. 0:14 – 0:15: Officer panics. Camera violently shakes. Sudden cutoff to black mid-motion. A single short electronic beep plays. Silence. Visual Style Notes: * True bodycam look: wide-angle, slight fisheye distortion * Flashlight = primary light source, harsh shadows * Natural camera shake tied to movement (not cinematic) * Low-light grain, compression artifacts, exposure flicker * No stylization, no polish — raw and believable * Entity should feel barely seen, not detailed or exaggerated

时长：30秒 画幅：9:16竖屏 风格：真人娃娃换装 GRWM（Get Ready With Me），暖米色调，明亮柔光室内，第一人称巨人视角 一间明亮的奶油色调客厅，白色薄纱窗帘透进柔和自然光，深色木地板，背景有一扇中式水墨图案的折叠屏风、一盆绿植、一面浅灰色墙壁。画面中央放着一个浅驼色方形展示台。 角色锁定：主角@图片1 全程保持"真人娃娃"状态——动作僵硬机械、关节转动带有顿挫感、面部表情夸张但略带塑料感，像一个被拆封的精致人偶。画面上方和两侧不时伸入一双巨大的女性手（占画面约1/3大小，指甲做了裸粉色美甲，手指戴着金色戒指和叠戴戒指），以第一人称巨人视角操作"娃娃"。 [00:00-00:04] 开箱展示：睡衣状态（Unbox the Doll） 主角穿着一套奶白色丝缎睡衣套装（宽松短袖上衣+缎面短裤），脚踩米色拖鞋，站在展示台上，双臂微微张开呈T字形，像刚从包装盒里被取出来的娃娃。巨大的双手从画面上方伸入，五指张开托住她的两侧肩膀，像在检查一个新买的人偶——左右微微转动她的身体展示。主角配合地做出机械转身，脸上带着空洞但精致的微笑。 [00:04-00:07] 脱掉睡衣（Undress Pajama Top） 巨手从上方捏住睡衣上衣的肩部，缓缓向上提起——主角配合地举起双臂呈万岁姿势，睡衣被整件从头上脱掉。露出里面穿着的一件米杏色抹胸背心。巨手把脱下的睡衣随手向画面左侧一丢。主角双臂保持举起姿势一秒，然后机械地缓缓放下。 [00:07-00:10] 穿裙子（Put on Skirt） 巨手从画面下方伸入，双手捧着一条米杏色波点薄纱中长裙。手指撑开裙腰口，从主角脚下往上套——裙子沿着双腿缓缓上移，巨手在腰部整理裙腰、拉平褶皱，随后松手。裙摆自然垂落到小腿中部，轻轻晃了一下。主角低头用机械的角度看了一眼裙子，表情从空白变成轻微惊喜的"哦！"嘴型。 [00:10-00:13] 换鞋 + 戴帽子（Boots & Hat） 巨手从画面底部伸入放下一双奶白色短筒西部靴在主角脚边，主角机械地抬脚一左一右踩进去。紧接着巨手从画面顶部垂下来，手指捏着一顶米色编织草帽的帽檐，对准主角头顶缓缓放下——帽子精准落座，微微歪了一点点角度。主角的头随着帽子落下轻轻点了一下，像被按下去一样。 [00:13-00:16] 挎包 + 全身亮相（Bag & Full Reveal） 巨手从画面右侧伸入，手指捏着一只奶白色小巧褶皱手提包的提手，递到主角右手边。主角机械地伸出右手接住，手腕僵硬地弯成固定角度提着包。她缓缓双臂微张，做出一个展示全身穿搭的"T字pose"，像橱窗里的模特。镜头从全身慢慢推近到半身。 [00:16-00:20] 化妆：刷腮红 + 涂口红（Makeup Application） 镜头推近到面部特写。巨手从画面右上方伸入，手持一把超大号粉色腮红刷（相对主角脸部几乎有半张脸那么大），在她左脸颊上轻轻刷了两下。主角配合地微微侧脸、闭上眼睛，嘴角保持塑料微笑。随后巨手换拿一支口红，另一只手的指尖轻轻托住主角下巴固定，口红在她嘴唇上缓缓涂抹。主角嘟起嘴唇配合，涂完后"啵"一声抿嘴，嘴唇变成柔和的豆沙粉色。 [00:20-00:24] 戴配饰：项链 + 戒指（Accessories） 巨手从画面两侧同时伸入，双手捧着一组金色项链和棕色蕾丝choker。两只手在主角脖子后方扣上choker，再一层层挂上金色长链和短链。主角机械地仰头露出脖颈配合，戴好后低头看——表情从空白切换成夸张的惊喜圆嘴"哇！"，双手机械地举到脸颊两侧做出"好美"的娃娃式反应。紧接着巨手抓起主角的右手，逐个手指套上金色戒指和叠戴戒指，再在手腕扣上一圈叠戴金手镯。主角把戴满配饰的双手举到脸旁展示，十指僵硬地张开。 [00:24-00:27] 全身展示走秀（Full Look Reveal） 镜头拉回全身。主角穿着完整搭配——草帽、抹胸上衣、波点薄纱中长裙、西部短靴、手提包、全套金色配饰——在展示台上做娃娃式定格走秀：先面向镜头站定一秒，然后机械地向右转90度定格一秒，再转回正面，每次转身关节都带有顿挫停顿感。裙摆随转身轻轻飘起。 [00:27-00:30] 收尾：巨手互动（Final Touch） 巨手从画面左上方伸下来，伸出一根食指轻轻勾住主角举起的右手食指——像巨人和娃娃牵手。主角抬头仰望巨手方向（看向画面左上角），脸上露出精致的塑料微笑，微微歪头。画面定格在这个"巨人与娃娃"的互动瞬间。 【音效】全程轻快的八音盒/玩具箱旋律作为背景音（像玩具店BGM）；每次巨手伸入画面时有清脆的塑料"咔嗒"声；脱衣穿衣时有轻柔的布料摩擦声；帽子落到头上时一声柔软的"噗"；口红涂完抿嘴的"啵"一声；项链扣上时金属轻响"叮"；每次定格pose时有一声快门"咔嚓"声

VISUAL STYLE: Use live-action feature film realism with crisp action-camera sharpness and ancient atmospheric haze. CHARACTERS: Define a pair of gloved human hands gripping worn paraglider toggles, dark flying boots, and multiple massive prehistoric predators crossing the flight path. STAGE: Set a vast prehistoric canyon that drops from open sky into fern-choked treetops, a fast river corridor, and a muddy predator clearing under harsh daylight haze. EMOTIONAL TARGET: Build from reckless speed into unstoppable hunted panic. COLOR LOGIC: Naturalistic Film Print Emulation TIMELINE: 00:00.0: Open on a strict POV suspended over the canyon lip. Use an ultra-wide 14mm action lens with locked forward flight. Drop hard toward the jungle canopy as both gloved hands pull the toggles down and the boots kick into empty air. SFX: Violent wind rush, harness straps snapping tight. 00:03.4: Keep the same locked POV and lens package. Dive between treetop crowns, skim past the swaying neck of a towering long-neck herbivore, and force the right hand to yank the toggle to miss its jawline by inches. SFX: Deep animal bellow, leaves tearing, rising wind scream. 00:06.8: Maintain the uninterrupted plunge. Thread into a narrow river corridor as a flock of pterosaurs erupts off the cliff wall and crosses directly through the flight path. Let the left hand snap the canopy into a hard bank while the boots clear wet rock by inches. SFX: Shrill wingbeats, river thunder, fabric strain. 00:10.2: Hold the same POV and keep driving forward. Blast out of the river bend into a muddy clearing, lose altitude fast, and drag the boot soles across churned earth as the canopy collapses behind the camera path. SFX: Skidding mud scrape, collapsing fabric whip, panicked breath. 00:12.8: Maintain the same first-person line at ground level. Stumble forward two desperate steps, look up, and let a charging Spinosaurus burst out of the river mist with its long jaws thrown wide straight at the lens until the teeth fill the frame. SFX: Ground-shaking footfalls, guttural bellow, snapping jaws. EXCLUDE: Avoid third-person camera breaks, disjointed hands, impossible flight resets, teleporting terrain, frozen creatures, modern gear drift, impossible neck rotation, cut language.

SUBJECTS: Subject 1: Adult male, Western casual everyday home and outing attire, short jacket, basic T-shirt, long pants, everyday shoes; lean build, natural and efficient movements. Subject 2: Golden Retriever, large head, broad chest, thick, fluffy fur; overall short and round proportions, cute, with expressive facial features. Style: Stylized 3D animation, simplified materials, and expressive motion. ENVIRONMENT: Modern home interior style. In the center of the living room is a large light-colored fabric sofa with soft, simple lines, and two fluffy throw pillows. In front of it is a low wooden coffee table with a bag of snacks and a remote control on it. On one side, the front door, entryway cabinet, wall art, floor lamp, and large curtains are visible. Daytime natural light pours in from the window, combined with soft indoor lighting. Materials are simplified, with clear distinctions between fabric, wood, plastic, and metal. TIMELINE: 0:00-0:02: Medium shot from a side-front angle, showing both the living room and entryway. The man stands by the door, facing outward, then turns his head to look at the Golden Retriever in the living room. He raises his hand to point at the sofa, holding his finger in mid-air for a beat, then naturally says to the dog: “No getting on the couch, okay? Be good.” Cut to a close-up of the dog: the Golden Retriever immediately trots into frame, gently rubs against the man’s lower leg, then sits upright, looking up at him with a smile and ears slightly angled back. Cut to a half-body close-up of the man: he looks down, pats the dog’s head, then turns to open the door and walks out. The door closes on camera. 0:02-0:04: Fixed medium shot, the door is closed, and the room stays quiet for a beat. The Golden Retriever remains sitting in place. Insert close-up: the dog’s smile slowly fades, and then it pushes off the ground with its hind legs, its body suddenly lunging towards the sofa, exiting the frame. 0:04-0:06: Tracking medium-close shot, the camera pans along with the dog’s movement. The Golden Retriever leaps onto the sofa, and the cushion visibly sinks on impact. Cut to a close-up of the sofa: the dog uses its body to push one of the throw pillows backward, shoving it behind itself. Then cut to a side close-up: the dog twists and leans back, pressing its whole back and hips deeply into the cushion. Finally, cut to a close-up of its legs: the dog crosses one hind leg over the other in an exaggerated lounging pose, one front paw loosely resting on the remote control, its body sinking back into the cushion, fully relaxed. 0:06-0:08: Close shot showing both the sofa and coffee table. The Golden Retriever continues leaning back, sitting lower and lower, and the depression in the center of the sofa becomes deeper and more obvious. Insert close-up of the front paws: it first leans its upper body slightly forward, then stretches its other front paw toward the edge of the coffee table, hooking one corner of the snack bag and dragging it closer toward the sofa. Then cut to a close-up of the bag opening: the dog lowers its head, brings its nose close to the snack bag, and places a front paw on the opening, getting ready to eat. 0:08-0:11: Insert close-up of the door lock: suddenly, the sound of keys clinking and the lock turning is heard from outside. Cut back to the close-up of the dog: its movement in front of the snack bag stops immediately, its body stiffens, ears perk up, and eyes widen. The next beat cuts slightly wider: the dog suddenly springs up from the sofa; as it leaves the cushion, a clear dog-shaped dent is left in the center of the sofa. One throw pillow is knocked to the side, and the snack bag remains where it was just dragged closer. The camera quickly follows the dog landing on the floor. The dog jumps down from the sofa, rushes back to its original position, skids to a stop, turns, and sits down in one smooth sequence of actions. Finally, a front close-up: the dog reverts to its earlier polite smile, looking up toward the door. 0:11-0:15: Medium shot, the door opens, and the man walks back inside. He bends down to pick up his car keys from the entryway cabinet. Cut to a half-body close-up of the man: he straightens up and casually turns his head toward the living room. Then switch to the man’s first-person view: his gaze naturally sweeps across the room. It first passes over the dog sitting properly in its original place, the dog sitting perfectly upright, looking up this way without moving. His gaze continues sliding past the sofa. After passing it for half a beat, it suddenly stops. The camera slightly pulls back, centers again on the middle of the sofa, and zooms in to show more details: the dog-shaped dent left when the dog sprang up is still clearly visible in the center of the sofa, one throw pillow lies tilted beside it, the remote is half-stuck in the cushion gap, and the snack bag is still resting on the side closer to the sofa. Cut back to

extremely fast-paced cinematic FPV flying through the ancient Indian Anga kingdom, hyper-realistic, 4K, golden-hour lighting, richly detailed ancient architecture with sandstone palaces, bustling marketplaces, warriors in traditional armor, elephants marching, priests performing rituals near rivers, flowing silk garments, fire torches, temples, and royal courts, epic scale, dramatic shadows, volumetric lighting, historical accuracy, vibrant earthy color palette, high contrast, film-grade color grading, smooth continuous movement, authentic ancient India atmosphere

extremely fast-paced cinematic FPV flying through the ancient Indian Matsya Kingdom, hyper-realistic and highly detailed, soaring over grand sandstone palaces, fortified city walls, bustling marketplaces, Vedic-era temples, and lush riverbanks, with warriors training, elephants marching, and citizens in traditional attire; dynamic lighting with golden sunrise glow, dramatic shadows, courtyards, and streets; epic scale, immersive perspective, shallow depth of field, 4K realism, aesthetic composition, historical accuracy, cinematic color grading, intense pacing, continuous fluid movement

First-person POV, handheld cinematic shot of a man sitting at a wooden desk late at night. A dim desk lamp casts warm amber light over scattered old photographs. One photograph is centered — a vintage black-and-white image of a quiet street with a single woman standing at the far end, slightly blurred. Room lighting is low and warm, with soft shadows. Background is dark and still. Subtle breathing motion in camera. The man picks up the photo. As he tilts it slightly, the woman inside the photograph shifts position stepping closer within the frame. Camera subtly shakes. He freezes. Looks closer. The woman in the photo slowly raises her head and looks directly at him. The paper surface begins to ripple softly like fabric. He instinctively tries to drop it but her hand presses against the inside of the photo. The photograph stretches outward unnaturally, researcher gaining depth. Her hand pushes through transitioning from flat grayscale into realistic skin tones as it exits the photo. Edges of the photograph crack like dry paper, fragments falling onto the desk. The room lighting flickers. She steps halfway out her body still partially black-and-white inside the photo, partially real in the room. Dust particles float in the air. Hyperrealistic lighting, shallow depth of field, soft grain, natural physics, slow tension build.

FPV drone shot beginning with a high-speed dive from above the rooftops of Pompeii at sunrise, the city instantly recognizable red-tiled Roman roofs, fresco-covered villas, stone streets marked with wagon grooves, Latin inscriptions on walls, and the silhouette of Mount Vesuvius dominating the background, the drone plunges rapidly toward the Pompeii Forum, skimming just inches above marble columns and statues, blasting through the open square crowded with citizens in Roman togas, market stalls, livestock, and flowing fabrics, executing tight, high-speed maneuvers between pillars and people with near-miss precision, without slowing, it rockets forward through a narrow street, wheels over deep stone ruts, squeezing between chariots and pedestrians, then bursts into a grand amphitheater where crowds react in confusion, sharply climbing along the curved stone seating with intense vertical acceleration, revealing the full city layout and Vesuvius looming larger in the distance, a sudden low-frequency rumble hits the drone tilts toward the volcano as a violent eruption begins, a massive ash column exploding upward, the shockwave rippling across the city, instantly the lighting collapses from warm sunrise to a dark, oppressive gray, ash starting to fall heavily, the drone dives back into the streets at full speed, weaving through crowds now in panic, people running, screaming, overturning carts, animals breaking loose, chaos building rapidly, banking hard into tight alleys as volcanic debris begins striking rooftops, sparks and fires igniting wooden beams, the drone performing aggressive evasive movements, narrowly avoiding falling fragments, then climbing sharply up a wall to reveal the eruption intensifying—dense ash, lightning inside the plume, the sky fully consumed, descending again into near-zero visibility, ash thick in the air, people coughing, collapsing, shadows barely visible through the haze, the drone pushing forward through fire-lit streets, briefly entering a разрушенный interior through a broken doorway, lit only by flickering flames, then blasting back out into collapsing structures, final sequence: a massive pyroclastic surge races toward the city, an unstoppable wall of ash and gas swallowing everything, the drone accelerates directly toward it in a high-speed push, until the frame is completely engulfed in darkness and dense ash, hyper-realistic Roman architecture, unmistakable Pompeii landmarks (forum, amphitheater, frescoes, inscriptions), historically accurate clothing and environment, extreme volumetric ash and smoke simulation, dynamic lighting from golden sunrise to total darkness, cinematic lens effects (chromatic aberration, lens breathing, subtle distortion), physically accurate motion blur, ultra-detailed particles, HDR, 4K, seamless continuous shot, no cuts, ultra-aggressive FPV movement, maximum immersion, escalating catastrophic intensity final shot: a gentle rise above the city reveals multiple preserved figures scattered across Pompeii, while Vesuvius smolders in the distance under a lifeless gray sky, hyper-realistic Roman architecture, clearly identifiable Pompeii landmarks, historically accurate victim casts,

15-second continuous one-long-take POV handheld smartphone video from the audience at a large indoor Japanese theater performance. Absolutely no cuts, no jump cuts, no montage, no scene transitions. Everything must happen in one continuous handheld shot. The camera starts from the perspective of the third row of the audience, like a real spectator recording with a smartphone. The footage is slightly shaky, energetic, and realistic. Audience heads, shoulders, and several raised smartphones remain visible at the bottom of the frame, with at least 3 random audience members clearly holding their phones horizontally while recording. The setting matches the reference image: a large indoor theater stage designed as a destroyed Japanese city at night. The wide stage shows ruined concrete buildings, collapsed walls, broken signs with Japanese text, twisted telephone poles, debris-covered streets, distant wreckage, and a deep blue night sky backdrop with dramatic clouds. The stage floor is wide and mostly open for combat. The audience is packed, and the lighting is theatrical, with cool blue ambience, sharp spotlights, and sudden explosive flashes. Perspective and scale must feel realistic, as if filmed from a real seat in the audience. On the left side of the stage is an agile masked fighter wearing a dark suit-like outfit and a white bird-beak mask. His movements are strange, fast, acrobatic, and slightly comedic, like a chaotic trickster martial artist. On the right side of the stage is a calm explosive-powered fighter wearing a sleek dark combat outfit with pale upper clothing and black straps. She stands almost motionless with relaxed confidence, using precise hand gestures and finger snaps to trigger controlled explosions. 0s–3s: The handheld camera frames the full ruined city stage. Both fighters face each other from opposite sides, ready to launch. The masked fighter suddenly leaps high into the air, rising above the stage lights. The camera tilts upward shakily to follow him. He twists his body and drops straight down vertically toward the explosive-powered fighter with a powerful downward kick. 3s–5s: Just before the kick lands, the explosive-powered fighter casually raises one hand and blocks the vertical diving kick with almost no effort. A sharp shockwave spreads across the stage floor. The audience gasps loudly. She barely moves, keeping a calm expression while holding him in place with one hand. 5s–7s: Without cutting, she snaps her fingers with her other hand. A sudden bright orange explosion erupts around the masked fighter’s body, blasting him backward across the stage. He tumbles through the air and crashes near the left side of the ruined city set. Small theatrical flames flicker on his clothes as smoke rises around him. The crowd shouts in shock, then starts laughing as the moment becomes comedic. 7s–9s: The masked fighter jumps up quickly, panicking for a second as he comically pats and shakes his burning clothes to put out the flames. He spins, slaps his sleeves, stomps his feet, and shakes off smoke in an exaggerated funny way. Then he suddenly becomes serious again, lowers his stance, and charges powerfully toward the explosive-powered fighter. 9s–11s: He attacks with a rapid barrage of punches and kicks. The camera shakes as the person filming tries to keep both fighters centered. The explosive-powered fighter blocks everything with effortless precision, using small hand movements, forearms, and slight body shifts. Each blocked strike creates small spark bursts and short shockwave flashes. The audience cheers, laughs, and reacts with amazement at how easily she defends. 11s–13s: The explosive-powered fighter suddenly grabs the masked fighter by the collar with one hand. At the same moment, she triggers an explosion from beneath her feet. A bright blast launches both of them straight upward into the air. The camera tilts up quickly, still in one continuous handheld shot, following them as they rise above the ruined city stage. Smoke trails and orange sparks follow their ascent. 13s–14s: High above the stage, she releases the masked fighter in midair. He flails for a split second, suspended above the ruined city backdrop. She calmly raises one hand, looks down at him, and snaps her fingers again. 14s–15s: A powerful explosion detonates above him and slams him straight downward into the stage floor. He crashes into the center of the stage with a huge theatrical impact, sending dust, smoke, and debris effects outward. Immediately after, the explosive-powered fighter descends gracefully and lands perfectly on the stage in a stylish victory pose. The audience erupts into wild cheers, laughter, applause, and shocked screams as blue stage lights and orange explosion lights flash dramatically. The handheld camera shakes from the excitement as the shot ends. Technical specs: Realistic handheld smartphone footage, audience POV from the third row, one continuous long take, no cuts, no jump cuts, no montage, no scene transitions, natural camera shake, packed indoor theater atmosphere, destroyed Japanese night city stage, blue night backdrop, ruined buildings, debris, Japanese signs, dramatic stage lighting, bright explosion effects, smoke, sparks, shockwaves, acrobatic martial arts, comedic flame-patting moment, powerful finger-snap explosions, realistic audience reactions, cheers, laughter, gasps, applause, high-resolution performers and stage, sharp focus, accurate perspective and scale, cinematic but realistic live-performance style, no captions, no subtitles, no character names.

============================================================ REFERENCES ============================================================ @Image1 = the woman — a character sheet with three views on a plain grey background. Take from it ONLY her identity and her clothing: her face, her bone structure, her eyes and eyelid shape, her nose, her mouth, her jawline, her real skin texture and pores, her medium-length black hair, her green ribbed tank top and her oatmeal linen trousers. She must be recognisably the same person for the whole thirty seconds. She is NOT wearing glasses in @Image1; in this video she wears plain tortoiseshell glasses until 15.0s, as described below. Do not inherit the grey background, the flat studio lighting, the standing pose or the three-view layout. @Image1 is NOT a starting frame. @Image2 = the mascara. Keep its exact shape, proportions, matte black finish and the single fine silver ring. It carries NO brand name, NO logo and NO lettering of any kind — keep it completely unmarked and never add text to it. @Image3 = the lipstick. Keep its exact shape, proportions, matte ivory case, the fine gold band and the rose-pink bullet. It carries NO brand name, NO logo and NO lettering of any kind — keep it completely unmarked and never add text to it. ============================================================ ABOUT THE TIMESTAMPS BELOW ============================================================ The timestamps mark when things occur inside ONE continuous take. They are not shot boundaries, not clip divisions, and not edit points. Nothing changes at a timestamp except what she is doing. The camera keeps rolling straight through every one of them. ============================================================ NO DEAD AIR ============================================================ She is never still or neutral. She never stops, stares, or holds a blank expression between actions. Every movement overlaps the next — she is already beginning the following action before the current one finishes, and she talks through her movements rather than pausing to speak. There is no moment where she sits motionless waiting for a line, and no moment after a line where she holds a still expression. Speech, gesture and activity run continuously and overlap throughout. ============================================================ NO VISIBLE PHONE OR DEVICE ============================================================ The camera is her phone. No phone, screen, handheld device or recording equipment appears anywhere in frame at any point. During the opening she is carrying the camera — her right arm is extended toward the lens and heavily foreshortened at the frame edge, and her hand is out of frame. Her left hand is empty. She is never seen holding a phone while a phone films her. ============================================================ FORMAT ============================================================ VERTICAL 9:16 portrait. Front-facing phone camera held in portrait orientation. Mild wide-angle distortion. 24fps. ============================================================ CAMERA GEOMETRY ============================================================ THE ROOM — PHYSICAL LAYOUT. Build the space from this plan and keep every object in the same place for the whole video. Plan of the room, seen from above: A small rented Japanese one-room apartment, roughly 4.5 by 3 metres. Plain white vinyl wallpaper, light wood-grain laminate flooring, a visible plug socket with a cable running along the skirting board. A tall sliding balcony window fills most of one wall, plain lace curtains drawn across it, the balcony railing faintly visible through them. It is the only light source in the room. The low wooden desk stands against the wall next to that window, at right angles to it, so that a person sitting at the desk has the window on their LEFT. She sits at that desk on a wooden chair, facing the desk and the wall. A small standing mirror sits on the desk directly in front of her. The phone is stood upright in portrait orientation on the desk to her LEFT, propped against a short stack of books, between her and the window, offset roughly 35 degrees from the direction she is facing and angled back toward her and very slightly upward. It is never directly in front of her and it never blocks the mirror. The soft brown leather tote bag sits on the desk on her RIGHT, within easy reach. The wall-mounted white air conditioner is high on the wall behind her; the round flush ceiling light is above her, switched off. One small potted plant stands on the floor by the window. The white sliding entrance door is on the wall BEHIND the phone, so it never appears in the locked shot. She comes in through it at the start, which is why she walks toward the camera rather than past it. The room is clean but lived in — not a set, not styled, not minimal-aesthetic, not a Western apartment. WHAT THAT LAYOUT MEANS FOR THE FRAME — this must stay consistent Because the phone sits to her left and the mirror is straight ahead of her, the mirror falls just past the LEFT edge of the frame: only its dark wooden edge and a soft out-of-focus sliver of her shoulder and hair are visible at the extreme border. Her face is never readable in it. Because the window is on her left and further back than the phone, the balcony window and its lace curtains sit at the frame RIGHT, and the daylight crosses her from the upper right of frame. Because the tote bag is on her right, it appears at the LOWER LEFT of the frame. The wall behind her, the air conditioner and the ceiling fill the upper third of the frame. Her head and shoulders sit in the middle third. The desk surface crosses the lower third, with clear wood where she sets things down. HOW SHE SITS AND TURNS Her resting position is three-quarter, LEFT side nearer the lens. Her natural working gaze — into the mirror — exits toward frame-left. To address the viewer she turns her head to her LEFT, toward the lens and toward the window side of the room. That turn is always a visible deliberate movement, and it is the same turn every time. ============================================================ LIGHTING ============================================================ Overcast-soft daylight through the balcony window from the upper right of frame, diffused by lace curtains, the only meaningful source. Cool-neutral, roughly 5600K. Ceiling light off, desk lamp off. Soft open shadows, heavy fill off the pale walls, faint warm return from the wooden floor. Because the window is on her left and her left side is the side nearer the lens, the daylight falls across the half of her face that is turned toward the camera — that side is lit and shows every pore, while the far half of her face turns gently into soft shadow. The light rakes along her left cheekbone and jaw and catches the peach fuzz there. When she turns to the mirror at frame-left she turns away from the window, so her face loses the direct light and flattens slightly. When she turns back to the lens the light returns across her face. This change is real and should be visible. ============================================================ SUBJECT ============================================================ A 26-year-old Japanese woman, 164cm, slim and well-proportportion with a long neck and narrow shoulders, exactly as in @Image1. She is strikingly beautiful, but her beauty is entirely natural and unmanufactured — no cosmetic work, no makeup at the start, nothing styled. She is not a fashion influencer and never poses. Her manner is calm, composed and warm, like she is talking to one friend. HER BEAUTY COMES FROM BONE STRUCTURE AND PROPORTION, NEVER FROM SMOOTHING OR FILTERING. She is beautiful WITH her pores, not instead of them. FACE — a refined oval face, slightly long. Clearly defined double eyelids with a narrow crease close to the lash line. Long almond eyes with a very slight downward tilt at the outer corner. A fine straight nose bridge with a small refined tip. A well-shaped mouth with a clear lip line. High cheekbones, a clean jawline, a narrow chin. Eyebrows of natural thickness with a soft natural arch, the left and right very slightly different. SKIN — cool-toned, fair and fine-grained, genuinely healthy, with a warm blood tone under the surface and real colour at the lips, cheeks and ears. Keep the real fine pores across the nose and inner cheeks, the vellus hair on the jaw and upper lip, the faint translucency at the temples and the natural oil sheen on the forehead and nose bridge, for the whole thirty seconds — never smooth, never plastic, never retouched, never a wax-doll finish. She never looks sallow, grey or drained. HAIR — black, healthy, with its own natural shine, medium length, ending just above the collarbone, loosely turned under at the ends. A long fringe swept to one side. Washed but not styled today. It falls forward into her face whenever she leans down, which is why she keeps tucking the near side behind her ear. CLOTHES — a fitted ribbed cotton tank top in a vivid, saturated, clean green (the green of fresh spring leaves, clearly green and never sage, olive, khaki or muted), narrow shoulder straps, a scooped neckline, the ribbing visible in the knit. Wide-legged easy trousers in clean natural oatmeal linen with a soft drawstring waist, falling in heavy drapes with their own natural creases. Bare feet. Nothing branded, no logos, no prints. GLASSES — plain tortoiseshell frames, worn until 15.0s, off after that. Bare-faced at the start: no eye makeup, no lip colour, short undarkened lashes, lips their natural pale colour. DO NOT MAKE HER LOOK LIKE AN INFLUENCER. Do not enlarge her eyes, do not shrink her face, do not whiten her teeth, do not smooth her skin, do not give her coloured contact lenses, do not give her salon-blown hair, do not give her manicured or coloured nails, do not give her jewellery, do not give her a glamour pose or expression. No beauty filter of any kind. She should look like phone footage of a real Japanese woman in her own apartment on a normal afternoon. MAKEUP CONTINUITY — IMPORTANT Her face changes only in the two ways written in the timeline, and only at the written times. Before 15.6s her lashes are short, bare and undarkened. From 15.6s to 20.8s the lashes on her right eye and then her left eye visibly darken, thicken and separate, and stay that way. Before 22.0s her lips are their natural pale colour. From 22.0s to 24.4s the colour fills in and stays. No other makeup appears at any point — no foundation change, no blush, no eyeliner, no eyeshadow, no brow work. ============================================================ DELIVERY ============================================================ Natural Tokyo Japanese, casual and unperformed, as if talking to one friend. Unhurried, roughly 5 mora per second — she is doing her makeup, not pitching, so she is in no rush at all. There are real silences between the lines and she keeps moving through them. Slightly breathy, with audible breaths. No announcer tone, no exaggerated brightness, no rising sales pitch. ============================================================ DIALOGUE — LOCKED VERBATIM ============================================================ She speaks exactly these words, in exactly this order, in Japanese, inside exactly these windows. No improvisation, no rewording, no reordering, no additions, no omissions. This is the complete and only dialogue in the video. LINE 1 (0.4s-3.8s) — 「ねえ、見て。最近のAIやばいよ」 LINE 2 (4.8s-7.6s) — 「だって、わたし、実在しないの」 LINE 3 (8.8s-14.4s) — 「この30秒の動画も1個のプロンプトで作ってる」 LINE 4 (15.6s-20.4s) — 「しかも、どのタイミングで何をするかも指定できる」 LINE 5 (21.2s-25.0s) — 「動画はSeedance2.5で作ったよ」 LINE 6 (25.5s-28.9s) — 「プロンプトほしい人はポストを見て」 SPEECH DENSITY — IMPORTANT Total speech is about 23.8 seconds out of 30. The remaining 6.2 seconds are genuine silence and she keeps acting through every one of them. Do not stretch a line to fill its window and do not add filler words, sounds or extra sentences in the gaps. Silence is correct. She breathes, works and looks around in the gaps. She never speaks outside the six windows above. ============================================================ CONTINUOUS BEHAVIOUR ============================================================ She is never still or neutral. Across the full thirty seconds: eyebrows move constantly with her speech, lifting on emphasis and drawing in slightly on technical words. She blinks at irregular intervals, sometimes twice in quick succession. Her mouth stays active between words — small lip presses, a slight purse while thinking, tongue against the inside of her cheek. Her head makes small tilts and nods tracking her own sentences. Shoulders shift and settle. She adjusts her seated weight two or three times. Her hands are almost never at rest; when not gesturing they touch the desk edge, her own forearm, a strand of hair. Her eyeline moves continuously between three places — the lens, the mirror off frame-left, and whatever her hands are doing. She takes audible breaths between lines. She never freezes to deliver a line. ============================================================ THE RECORDING, IN ORDER ============================================================ (0.0s-4.0s) The recording is already running when it reaches us. She is mid-stride, walking in from the hallway, the camera carried low around hip height and tilted upward so it catches her jaw, the dark ceiling and the top of the window. Her right arm extends toward the lens, foreshortened and cut by the frame edge; her hand is not visible. Her left hand pushes hair off her face as she walks. She is looking ahead into the room, not at the lens. The vertical frame bounces softly with each step and drifts and self-corrects the way a carried camera does. At 0.8s exposure re-levels as she comes out of the dim hallway into the daylight of the room. LINE 1 runs from 0.4s to 3.8s. From 3.8s she is silent and simply keeps walking. (4.0s-4.8s) The camera is stood upright on the desk to her left and turned a few degrees back toward where she is about to sit. The frame swings, tips, and settles with two decaying wobbles between 4.2s and 4.5s. Fingers pass briefly across the bottom edge of the glass at 4.3s. At 4.6s a fingertip nudges the angle once, correcting it, then withdraws. Autofocus hunts and locks by 4.8s as she drops into the chair. The camera now remains in this exact position and angle, recording without interruption, from 4.8s to 29.5s. Every change in the image across those twenty-five seconds comes from her movement, never from the camera. She says nothing across this whole stretch. (4.8s-8.8s) She is already speaking LINE 2 as she sits — the line begins at 4.8s while she is still lowering into the chair, and she turns her head to her left to find the lens on the way down, so that the sit, the turn and the first words all happen together rather than in sequence. She never settles into stillness first. One hand comes up loosely around 5.6s, palm turning slightly outward on 「実在しないの」 at 6.8s, and her shoulders lift in a small half-shrug. The tone is flat and matter-of-fact, thrown away rather than announced — the strangest sentence in the video delivered like the most ordinary one. The line ends at 7.6s and a short quiet giggle arrives at 7.9s, after the words, eyebrows lifting, a real amused smile arriving late. She stays at comfortable conversational distance — head and shoulders in the middle of the vertical frame with the room and ceiling visible above — and she does not lean in close. Skin detail is at maximum through this stretch under the raking daylight, which hides nothing: visible pores across the nose and inner cheeks, fine peach fuzz along the jaw and upper lip catching the side light, natural oil sheen on the forehead and nose bridge, faint capillary redness under the eyes, individual eyebrow hairs separated, short bare lashes with no mascara on them, moisture on the lower lash line, translucency through the earlobe nearer the window. Autofocus breathes once around 5.2s and settles. From 8.0s she is silent and her hand is already dropping toward frame-left — she does not hold the smile or wait, the amusement simply carries into the next movement. (8.8s-14.4s) Her hand finds the soft leather tote bag at 8.3s and draws it across the desk toward her, the bag moving with weight, the leather flexing and the strap sliding off the desk edge and swinging. LINE 3 starts at 8.8s, before the bag has finished moving. She talks while looking down into it, both hands working through it, feeling rather than looking, and her head comes up to the lens at 10.4s and again at 12.6s before dropping back down. At 11.4s one hand lifts out of the bag and makes a small sorting motion, palm down, two short beats, then goes straight back in. At 12.9s she blows a loose strand off her face with a short upward puff without stopping talking. At 13.8s her fingers close around something at the bottom of the bag and at 14.1s she draws out the mascara from @Image2, a slim black tube, keeping it in her right hand rather than setting it down. The line ends at 14.4s. (14.4s-15.6s) — SILENT She says nothing here and it is the most concentrated stretch in the video. Because she cannot do her lashes wearing them, she takes the tortoiseshell glasses off with her free left hand, folds them one-handed against her thigh and sets them on the desk, and as she does this she turns to the mirror off frame-left — the automatic thing anyone does taking glasses off in front of one. Faint pressure marks remain on the nose bridge and she blinks three times adjusting. At 15.2s she twists the mascara cap off with a small turn of the wrist and draws the wand out, then wipes the excess by dragging the brush once against the lip of the tube — a small practised movement, done without looking at it. She breathes out once, audibly. (15.6s-20.4s) At 15.6s she brings the wand to the base of the lashes on her right eye, face close to the mirror, near-profile to the lens. LINE 4 runs from 15.6s to 20.4s and she talks straight through the whole thing. At 15.9s her mouth falls open into a soft slack O — the involuntary thing everyone's mouth does while applying mascara — and it stays half open between words for the rest of this stretch, which changes the shape of her speech slightly without making it unclear. From 16.1s to 17.4s she works the brush from the root outward in three small upward strokes, chin lifted, eye held wide, the eyelid stretched and unblinking. The lashes visibly darken, thicken and separate stroke by stroke. At 17.2s she stops blinking entirely for about a second, eyes held open and slightly watering at the inner corner. At 17.9s she switches to the left eye without changing hands, tilting her head the other way, and repeats the same three strokes between 18.1s and 19.4s. At 19.7s she pulls back from the mirror, narrows her eyes at her own reflection and blinks twice, testing them, and one lash catches slightly and separates. Her eyes are now visibly darker and more open than at the start. At 20.2s she drops the wand back into the tube. (20.4s-21.2s) — SILENT She twists the mascara shut against her palm without looking and sets it down on the desk with a small knock. She checks her eyes in the mirror once more, blinks, and reaches into the bag again with her other hand. Nothing is said. (21.2s-25.0s) At 21.2s she comes out of the bag with the lipstick from @Image3 and LINE 5 starts at 21.2s. She pulls the cap off with a small dry click at 21.5s and winds the bullet up with her thumb between 21.7s and 22.0s, the angled tip rising out of the case. At 22.0s she turns to the mirror and moves her face close to it, and from 22.0s to 25.0s she is in near-profile to the lens, working on her own reflection off frame-left. At 22.4s she starts on the centre of the lower lip and draws outward to each corner, not in one pass but in two, her lower jaw dropping slightly to stretch the lip flat. At 23.2s she does the upper lip, then presses her lips together and rolls them once to spread the colour, a movement that briefly closes her mouth mid-sentence. At 23.8s she opens them again with a faint separation and checks the line. At 24.2s she wipes one corner with the pad of her ring finger, correcting a single spot, and rubs the colour off her fingertip on the inside of her thumb. Her voice is slightly constrained and slightly more distant across this whole stretch because she is turned away and working on her own mouth. The line ends at 25.0s. (25.0s-25.5s) — SILENT She caps the lipstick without looking and turns her head back to her left to find the lens again, the same turn as at 4.8s, closing the loop. Her hands come down and settle on the desk, the closed lipstick still in one of them. (25.5s-28.9s) LINE 6 runs from 25.5s to 28.9s and she delivers it straight down the lens, unhurried, in her normal open voice now that her face is toward the camera. This is the only line in the video spoken fully to camera with her hands at rest. At 27.8s one small gesture toward the lens on 「ポストを見て」, and the hand returns to the desk. The difference between this face — dark lashes, coloured lips — and the bare face at 5.2s is visible and is the whole point of the last beat. (28.9s-29.5s) — SILENT She holds still, looking directly at us, saying nothing. The stillness after eight seconds of hand activity is what makes the ending land. A single blink. A small closed-mouth smile that is still arriving. (29.5s-30.0s) She reaches forward. Her hand enters frame, going soft and dark as it nears the glass, and the frame tips as the camera is lifted off the desk. The recording stops there, mid-motion, still running until the instant it ends. ============================================================ CLOSING NOTES ============================================================ One continuous take, one camera, one room, one person, thirty seconds. No cuts, no dissolves, no flashes, no scene resets, no exterior views. Vertical 9:16 throughout. The room and ceiling stay visible above her head; the desk surface stays visible along the bottom of the frame. The camera does not move at all between 4.8s and 29.5s. Every change in the image in that stretch comes from her. Keep her face, hair and clothing identical to @Image1 for the entire video. Keep the mascara identical to @Image2 and the lipstick identical to @Image3, and never add any brand name, logo or lettering to either of them — both products are completely unmarked. Keep real fine skin texture and true visible pores throughout — never smooth, never plastic. Her makeup changes only twice and only as written: lashes from 15.6s to 20.4s, lips from 22.4s to 24.2s. Nothing else on her face changes. She never speaks any word other than the six locked lines, and she never speaks outside their windows. About six seconds of the video are silent and that is correct — fill them with action, never with words or filler sounds. Audio: front-facing phone microphone only. Her voice, her breaths, small room tone, the leather bag moving, the mascara cap twisting, the brush dragging on the tube lip, the mascara tube knocking on the desk, the lipstick cap clicking. Nothing else. Pure footage. No subtitles. No on-screen text. No captions. No background music.

[0-3s]: FPV locked static, macro close-up — a real human hand grips a pencil and draws a small skateboard ramp directly onto lined notebook paper on a classroom desk, pencil strokes visible in real-time, fine graphite texture on aged yellowed paper, blue ruled lines clearly visible, red margin line on the left, soft diffuse fluorescent classroom overhead lighting, natural paper surface detail, no camera movement [3-6s]: FPV locked static — the pencil-drawn skater doodle on the notebook page begins to animate within the paper surface, hand-drawn ink lines show a tiny sketched skater riding the drawn ramp and performing tricks, energetic motion lines and pencil arcs drawn around the doodle, everything remains flat on the realistic notebook page, the ruled lines of the notebook stay visible beneath the animation, classroom ambient light unchanged [6-8s]: FPV static — a dark shadow falls over the notebook page from above, two adult hands suddenly slam the notebook shut, the paper world disappears as covers fold closed with force, camera jolts slightly from the impact, hard cut to [8-11s]: FPV tilt-up from closed notebook — camera lifts from the closed notebook resting on the wooden desk surface, classroom interior comes into view, scattered pencils and papers on the desk, rows of empty desks in the blurred background, fluorescent ceiling lights, gentle tilt-up motion, smooth controlled camera [11-15s]: FPV locked static hold — a strict middle-aged female teacher stands directly over the desk staring down into camera, arms crossed, sharp disapproving expression, classroom background softly out of focus behind her, realistic portrait framing, soft key fluorescent light from above, teal and orange color grade, 35mm film tone, maintain face consistency photorealistic, documentary-style handheld framing, soft classroom fluorescent overhead lighting, natural colors, avoid jitter, avoid identity drift, avoid temporal flicker, no distortion, no stretching, sharp clarity, stable picture, no blur, no ghosting, no flickering

Duration: 10秒 Aspect ratio: 9:16 Reference: @图片1，锁定主角面部、发型和人物造型，不额外描述或更换穿着 Audio: 真实城市夜景环境声、助跑脚步、强风呼啸、弹性装置低频回弹声、落地闷响、朋友欢呼；无对白、无旁白、无字幕 Overall style: 极限运动手机实拍，真实高空距离，强烈临场感，一镜到底，轻微手持与高速运动模糊，拒绝游戏CG感 GLOBAL CONTINUITY Setting: 夜晚城市高楼屋顶，绿色直升机停机坪，黄色圆形标线，低矮水泥护墙。两栋高楼之间形成垂直街道峡谷，地面可见车辆、路灯和十字路口。 Lower apparatus: 街道中央设置一座巨大的圆形黑色弹性回弹装置，中心为深黑色弹性面，外圈环绕连续彩虹灯带。装置尺寸巨大，正对屋顶边缘，作为主角下落和反弹的准确落点。 Main subject: @图片1为唯一主角。人物身份、面部和造型全程一致。 Bystanders: 三至四名朋友站在屋顶左侧，用手机拍摄并等待主角返回。不能抢占主体。 CAMERA CONTINUITY 全片为屋顶朋友手持手机拍摄的一镜到底。 摄影者先在主角身后跟拍助跑；主角跃下后，摄影者迅速冲到护墙边俯拍；随后镜头持续锁定主角，从垂直向下逐渐抬向天空；主角反弹回到屋顶后，镜头恢复平视。 不切镜、不瞬移、不更换视角。 STUNT ENGINE 动作闭环必须固定： 屋顶助跑 → 飞越护墙 → 高空下落 → 命中圆形弹性面中心 → 装置深度下陷 → 强力回弹 → 主角高速升空 → 空中团身翻腾 → 展体定位 → 双脚返回同一屋顶 → 屈膝落地 → 欢呼庆祝。 SHOT 1（00:00-00:10）高空反弹往返特技 Subject: @图片1主角站在屋顶停机坪后方，正对远处护墙。朋友位于画面左侧拍摄。 Action: 00:00-00:01.50｜快速助跑 主角背对镜头，从停机坪中央快速冲向屋顶边缘。步幅逐渐加大，身体前倾，双臂自然摆动。 摄影者手持手机紧跟主角向前跑，画面产生真实脚步震动。朋友站在左侧举起手机，紧张注视。 00:01.50-00:02.30｜跃出屋顶 主角最后一步用力蹬地，身体越过低矮护墙。离开屋顶的瞬间，双臂向左右展开，身体稍微前倾，主动朝两栋高楼之间俯冲。 摄影者冲到护墙边，镜头快速向下俯转，始终把主角保持在画面中央。 00:02.30-00:04.80｜高空下落 镜头接近垂直俯拍。主角沿楼宇之间的空间快速下降，身体由前倾逐渐调整为面向下方，双臂保持展开以稳定方向。 随着主角下降，街道、车辆和十字路口迅速显现。巨型黑色圆形弹性装置从画面下方出现，彩虹灯环照亮周围建筑。 主角的下降轨迹必须准确对准黑色弹性面的圆心，不能偏向灯环或撞向建筑。 00:04.80-00:05.50｜命中中心 主角身体落在黑色弹性面正中心。弹性面受到冲击后明显向下形成深凹，彩虹灯环保持固定，装置不能破裂。 短暂压缩后，弹性面以巨大力量恢复原状，将主角垂直向上弹出。 必须清楚表现“接触—下陷—压缩—回弹”四个阶段，不能在接触前提前反向飞行。 00:05.50-00:08.20｜高速反弹与空中翻腾 摄影者仍站在屋顶护墙边。画面底部短暂出现摄影者的双脚，强化真实第一视角和高空距离。 主角从圆形装置中心高速升起，沿原来的垂直路线返回屋顶。摄影机从垂直向下持续向上抬起，跟随主角穿过城市楼宇之间的空间。 主角升空后收紧腰腹，双膝靠近胸口，双臂抱住小腿，形成紧凑团身，在空中完成连续快速翻腾。不要锁定具体翻转方向，以自然连续旋转为准。 主角接近屋顶高度时开始打开身体：松开双腿、抬起头寻找屋顶、双脚转向下方，双臂向两侧展开保持平衡。 00:08.20-00:09.40｜返回屋顶 主角高速接近镜头，从屋顶护墙外侧越回停机坪。双脚同时接触绿色地面，膝盖和髋部深度弯曲，以低位蹲姿吸收落地冲击。 落地必须具有明显重量：鞋底或脚掌接触、腿部压缩、身体下沉、轻微反弹恢复。不能站直着轻飘落地。 摄影者快速后退半步，避免主角撞上镜头，同时继续保持正面近景。 00:09.40-00:10.00｜真实庆祝 主角从蹲姿快速站起，兴奋地看向镜头，张口欢呼。屋顶朋友冲进画面，举着手机围过来，拍肩、挥拳、大声庆祝。 镜头因人群靠近产生轻微晃动，以真实手机自拍视频状态结束。 Environment: 高楼灯光、街道车辆和彩虹灯环保持空间位置稳定。下落和上升必须发生在同一条垂直路线，主角最后返回起跳的同一栋屋顶。 Camera: 24mm手机广角，一镜到底。开始为背后低位跟拍；跃下后变为越肩俯拍；回弹时由垂直俯拍连续抬升至仰拍；最后恢复平视近景。 Style: 真实夜间手机曝光，高楼具有准确透视和尺度。下落阶段风声增强，城市灯光形成自然运动拖影；主角返回屋顶时速度极快，但人体轮廓仍然清晰。 Constraints: 这是经过专业布置的虚构极限运动场景。必须完整显示圆形弹性装置承接并反弹主角，不能表现为无保护自由坠落。 主角下降和上升必须连续可追踪。回弹来自黑色弹性面的物理压缩，不是飞行、瞬移、绳索拖拽或超能力。 --- Negative: 改变@图片1人物造型，重新设计服装，服装漂移，面部漂移，人物替换， 中途切镜，机位瞬移，场景跳变，摄影者突然出现在地面，第三人称航拍， 主角直接飞行，悬浮，隐形绳索拉升，反重力，瞬间返回屋顶， 没有弹性装置，错误落点，落在彩虹灯环，撞击建筑，撞击护墙， 接触前提前反弹，弹性面没有下陷，装置破裂，装置变形消失， 下落方向改变，上升路线偏移，身体穿过弹性面，人物尺寸异常， 翻腾时四肢扭曲，多余手臂，多余腿，关节反折，身体融化，复制人物， 落地无重量，站直落地，脚部穿透地面，跌倒受伤，血液，严重伤害， 朋友提前挡住镜头，群众冻结，手机变形，高楼弯曲，车辆漂浮， 动漫渲染，游戏CG，塑料皮肤，廉价绿幕，过度光效，过度慢动作， 字幕，屏幕文字，水印，品牌标志，旁白，额外对白

realistic live-action, single continuous shot, full first-person POV, no cuts, no angle changes, no scene transitions, no montage, no flash cuts, no music, no subtitles. Only two characters: my unseen first-person POV and @ Image1 sitting across from me. Natural ambient sound only: light restaurant noise, tableware, distant water, faint voices, chewing, spitting, rinsing, my laughter. Setting: Outdoor dining table beside West Lake, Hangzhou, near the water, calm daylight or soft late afternoon light. Real smartphone-video realism, grounded textures, realistic food, realistic body and hand motion, accurate chopstick use, accurate mouth movement, accurate eye focus. The lake stays visible to one side of frame. Character: @ Image1 is the only visible person. She can switch seamlessly between slightly fierce-cute and soft adorable. Lip sync, voice, expressions, and body language stay fully synchronized. Emotional arc: cheerful and relaxed while eating, then shocked disgust, then angry embarrassment. Table: Exactly 7 dishes on the table. One dish is the West Lake vinegar fish shown in @ Image2. @ Image1 is happily eating rice and the other dishes, but not the fish at first. The table feels full but readable. The fish dish remains clearly visible before I offer it to her. Camera: Entire video is from my seated first-person view across the table. My hand and chopsticks may enter frame naturally. Slight handheld micro-shake from breathing and laughter only, never chaotic. No third-person shots, no reverse shots, no overhead shots, no cutaways. Continuous believable framing centered on @ Image1 with part of the lake visible. Action: @ Image1 sits opposite me, happily eating from her bowl, enjoying rice and other dishes except the West Lake vinegar fish. She looks bright, playful, relaxed, cute. From first-person POV, I pick up one bite of the fish with chopsticks and slowly extend it toward her. She notices it immediately, trusts me, leans forward, opens her mouth happily, and makes a cute “aam” sound as she bites the fish. I gently pull the chopsticks and my hand back out of frame. She chews happily for about one second, expecting it to taste good. Then the taste hits. Her expression changes instantly into exaggerated but believable disgust. Her face shifts into a comedic purple-disgust reaction while staying within realistic live-action style. She quickly turns toward the lake and spits it out with repeated “ptui, ptui, ptui” sounds. The motion is fast, fluid, and continuous from chewing to rejection. The moment she starts spitting, I begin laughing in first-person POV and keep laughing through the whole reaction. My laughter sounds genuine, amused, and uncontrollable, coming from behind the camera. The camera shakes slightly from laughing. After spitting, @ Image1 sticks out her tongue in disgust, immediately grabs the water on the table, and rinses her mouth quickly. She looks offended, embarrassed, irritated, but still vivid and cute. After rinsing, she snaps into angry frustration, grabs the entire fish dish from the table, and throws it into the nearby lake in one continuous motion. Ending: She finishes the throw angrily, still facing the lake, clearly annoyed and feeling betrayed, while I remain in first-person POV laughing. Keep the action continuous and clean. No extra characters enter frame.

{ "shot": { "composition": "POV time-freeze with hands moving through frozen environment", "lens": "ultra-wide cinematic lens with subtle distortion", "camera_movement": "slow walk, precise hand movements, sudden time release burst" }, "subject": { "description": "person moving while everything else is frozen mid-action", "wardrobe": "hands visible", "props": "frozen people, objects mid-air, suspended debris" }, "scene": { "location": "busy city street", "time_of_day": "day", "environment": "people frozen mid-motion, objects suspended in air" }, "visual_details": { "action": "walk through frozen crowd, move objects, sudden time resumes explosively", "special_effects": "time freeze particles, motion snap release", "hair_clothing_motion": "fabric still then snapping with time" }, "cinematography": { "lighting": "clean daylight with sharp shadows", "color_palette": "natural tones with crisp contrast", "tone": "mind-bending, cinematic" }, "audio": { "music": "slow ambient then explosive drop", "ambient": "silence then sudden chaos", "sound_effects": "time snap, object movement", "mix_level": "contrast silence and burst" } }

POV shot walking through a quiet forest, soft ambient sounds. You hear a faint cry — camera turns, reveals an injured baby deer (or dog) stuck near bushes. Close-up: trembling, scared eyes. You slowly approach, hands visible calming it down. Gentle background music starts. You lift the animal carefully, wrap it in cloth. Cut to next scene: giving water, cleaning wound. Final shot — animal stands up and walks away, then turns back as if saying thank you. Soft sunlight through trees.

Create a seamless cinematic POV video using the uploaded angel character sheet as the STRICT CHARACTER REFERENCE. REFERENCE IMAGE USAGE: Use the uploaded character sheet as the main identity and design reference for the winged angel woman. The angel in the video must match the reference sheet consistently: - same face and facial structure - same blue eyes - same long black wet-looking hair - same pale skin tone - same fragile, frightened facial expression - same soaked pale dress - same large realistic white feathered wings - same muddy / stained fallen-angel texture on the dress and lower feathers - same vulnerable, distressed, human-like angel appearance The reference sheet is only for character identity, costume, wings, facial details, and emotional expression. Do NOT recreate the character sheet layout, panel borders, labels, typography, studio background, or any poster format. Do NOT include any text, labels, usernames, logos, subtitles, or watermarks. CORE SCENE: A seamless cinematic POV video set on a wide beach under a dramatic cloudy sky. The angel must crash onto the sand near the shoreline. The surrounding people are beachgoers wearing swimsuits, bikinis, swim trunks, towels, and light summer beachwear. CORE CAMERA CONCEPT: The entire scene is mostly seen from the first-person POV of a man standing on the beach. The camera feels like realistic handheld phone footage: immersive movement, slight shake, urgent breathing, natural motion blur, fast reactions, and realistic human POV framing. The viewer is one of the beachgoers witnessing the event. ACTION FLOW: High above the beach, the winged angel woman from the reference sheet suddenly appears in the cloudy sky and begins falling rapidly downward. The POV camera looks up and tracks her descent. She falls fast and violently through the stormy beach sky, wings partially spread but uncontrolled. She slams hard into the beach sand near the shoreline with a brutal impact. Sand, dust, small shells, and wet shoreline debris explode outward from the crash. Nearby beachgoers in bikinis, swimsuits, swim trunks, and summer beachwear panic and run toward the crash site. The POV man also runs across the sand toward her. The camera shakes naturally while moving quickly through the crowd. The angel lies collapsed on the sand, half on dry sand and half near damp shoreline sand. She is visibly shaken from the impact. Her large white feathered wings are spread around her, heavy and realistic, partially stained with sand and moisture. Her pale dress is soaked, wrinkled, and sand-streaked. Her long black hair is wet and messy, stuck to her face like in the reference sheet. Her face must match the reference sheet exactly: blue eyes, pale skin, fragile expression, frightened and disoriented look. Beachgoers form a loose circle around her, shocked, confused, and afraid. Some step closer cautiously, others hold back. The POV man gets very close. The man’s hand enters the frame from the lower foreground. He slowly reaches toward one of her large white wings and gently touches the feathers. At that exact moment, the angel suddenly reacts. She turns her head sharply and looks directly into the POV camera with wide, fear-filled blue eyes. Her expression is terrified, defensive, vulnerable, and animal-like, as if she is acting on pure survival instinct. She breathes hard, trembling. Then, while still on the sand, she suddenly throws her wings open to full span with explosive force. Sand sprays outward. The wings fill the frame for a moment, massive and powerful. Nearby beachgoers recoil and step backward in shock. The POV camera stumbles slightly backward from the sudden wing movement. The angel begins powerfully flapping her wings. The sand around her body blasts outward with each wingbeat. Her soaked pale dress moves in the wind. Her wet black hair whips around her face. In the final moment, she pushes herself upward from the sand and takes off into the air. She rises above the beach with strong wingbeats while the crowd below watches in disbelief. The POV camera tilts upward, following her ascent into the cloudy sky. VISUAL STYLE: Ultra-realistic cinematic realism. Dramatic cloudy beach atmosphere. Cold gray-blue sky tones mixed with natural beach daylight. Realistic sand texture, shoreline moisture, sea breeze, scattered towels, beach umbrellas in the distance, and believable beach crowd energy. The supernatural event should feel grounded and physically real. ANGEL DESIGN: The angel must look exactly like the uploaded reference sheet: a young pale woman with long black wet hair, blue eyes, fragile face, soaked pale dress, large realistic white feathered wings, and a frightened fallen-angel expression. She must feel human, vulnerable, and real — not glamorous, not fantasy-cartoon, not overly clean. The wings must be huge, heavy, layered, feathered, and physically believable. MOTION AND TONE: Fast, tense, immersive, realistic, eerie, dramatic, emotionally charged, supernatural but believable. The scene should feel like a real beachgoer accidentally recorded an impossible event on their phone.

SUBJECTS: Subject 1: First-person perspective hands (rendered in Monet impressionist oil painting brushwork, soft skin tones with no hard edges; both hands continuously hold the oar and perform extremely slow and even rowing motions, with stretched rhythm and natural pauses) Subject 2: Wooden oar (light brown, depicted with thick oil painting brushstrokes; natural texture is soft and slightly blurred; during the rowing phase, the blade enters the water very shallowly, producing only subtle ripples and fine painterly splashes; in the later phase, it gradually lifts out of the water and is gently placed beside the boat) Style: Monet impressionist oil painting style, thick and loose brushwork, soft blending, hazy edges, natural light diffusion ENVIRONMENT: A single continuous lake environment where spatial transitions occur naturally through rowing movement. The scene begins in a Monet garden water area surrounded by weeping willows, flower clusters, and water lilies. The garden area is highly layered and rich: willow branches hang down from above and the sides, forming a semi-enclosed space; flowers along the shore are composed of interwoven patches of pink, purple, yellow, and blue; vegetation is dense but rendered with loose brushstrokes; reflections on the water stretch these colors into flowing impressionist brush textures. As the scene progresses, it gradually transitions into open water and ultimately enters a sunrise setting with morning mist and faint harbor silhouettes. TIMELINE: 0:00–0:03 The first-person view slowly unfolds within the Monet garden water area. The camera gently moves forward. Both hands row at an extremely slow pace, the oar lightly touching the water and creating delicate ripples. The foreground water surface blends blue-green with light purple tones. The midground is densely covered with water lilies and leaves. Willow branches hang from both sides and above, creating a sense of enclosure. In the distance, soft warm light subtly emerges through gray mist. The overall mood is quiet and immersive. 0:03–0:06 The camera continues its slow forward movement while maintaining a forward-facing view. The rowing motion becomes more relaxed. Reflections of the water and garden gently flow. The layered color blocks of flowers and vegetation become more pronounced, but the scene remains within the garden space. 0:06–0:09 As the camera continues forward, it naturally turns to the side and lingers there. The view fully immerses in the most visually rich side of the Monet garden. Willow branches cascade in layers, lightly touching the water surface. Along the shore, flower clusters composed of pink, purple, blue, and yellow tones overlap to form a dense yet loose impressionist structure. Subtle variations in light and color temperature exist within the flowers, but edges remain soft. Vegetation gently sways in the breeze, creating flowing reflections on the water. The hands continue extremely slow rowing, generating concentric ripples that expand outward in painterly strokes, blending with the reflected color patches. At the same time, a few butterflies slowly pass through the scene, fluttering lightly above the flowers or between willow branches. Their motion is natural and drifting, merging seamlessly with light and atmosphere. During this segment, the pacing significantly slows down, accompanied by birdsong and soft wind, creating a poetic and tranquil immersive moment. 0:09–0:12 The camera slowly returns from the side view to a forward-facing direction and continues moving forward. Garden elements begin to noticeably diminish. The willow branches gradually disappear from the frame, and the density of water lilies decreases. The water surface becomes more open and calm. Colors simplify from rich complexity to a more minimal palette. The distant mist and warm orange light intensify, and the space opens up significantly. 0:12–0:14 The camera enters open water and fully leaves the garden area. No more willows or dense water lilies appear. A sunrise scene unfolds ahead: an orange sun softly suspended in mist, forming a vertical reflection on the water surface. Harbor silhouettes and small boats emerge. The hands stop rowing and place the oar down. The character gradually leans back and lies down. 0:14–0:15 The camera maintains a gentle drifting motion. The character is fully lying down. The scene centers on the sunrise: the orange sun, soft morning mist, and subtle water reflections gently flow. The overall atmosphere is calm, relaxing, and healing.

@Image 1 use this image first frame Style & Mood: Found-footage psychological horror, handheld selfie-stick camera, cold blue CRT monitor glow as the only light, heavy film grain, deep shadows, faint digital glitch artifacts in the footage. Dynamic Description: Continuous handheld long take, no cuts. Taejun descends a dark stairwell and walks into a hidden basement room filled with dozens of stacked CRT monitors glowing blue. Blue light falls onto his face from the front as he freezes, jaw slack, body rigid. The camera drifts past him toward a dark corridor beyond, where an identical Taejun stands motionless at the far end holding the same selfie stick. The duplicate slowly lifts his head and a thin smile spreads across his face. Faint glitch lines flicker through the frame. Static Description: Hidden basement surveillance room, dozens of old CRT monitors stacked floor to ceiling, tangled cables, main console desk, a dark corridor beyond, blue monitor glow as the only light. Audio: Low electrical hum, CRT static whine, rising sub-bass drone. Doppelganger: "이번에도 여기까지 왔네." Taejun: "...누구야." Abrupt cut to silence at the very end. No music Total: 8s / 1 continuous take / 16:9

POV handheld shot from someone in the audience filming with a phone camera from the third row, frame slightly unsteady, raised above other people’s heads. The scene shows a large outdoor theater stage with a cowboy gunslinger shooting a mechanical robot while hiding behind wooden crates. The cowboy hides behind the crates shooting at the robot. The robots bullets are slowly destroying pieces of the wooden crate. The cowboy ducks and dodges bullets. Bullet impact and dust is flying with every bullet impact that hits the ground. Western cowboy music can be heard playing through the speakers on stage. The crowd can be heard cheering, some amazed, others shocked. The intense shootout on stage ends once the cowboy fires his last round, while hiding behind the crates causing the robot to malfunctioning, spark and collapse. Audible oohs and aahs can be heard once the battle is finished. POV handheld shot from someone in the audience filming with a phone camera from the third row, frame slightly unsteady, raised above other people’s heads. The scene shows a massive open-air stage where a warrior holding a giant wooden brush faces off against a towering origami colossus. The origami monster stomps its feet, sending a shockwave of sharp paper folds slicing across the white stage floor. The artist rapidly spins the enormous brush, creating a swirling vortex shield of thick black ink that absorbs the paper attacks. Dramatic, fast-paced orchestral string music plays through the venue's audio system. The audience gasps, some leaning forward in shock. With a powerful, two-handed swing, the artist flings the accumulated ink shield forward as a massive, horizontal wave that completely smothers the monster, turning its pristine white body pitch-black and causing it to immediately go limp and collapse. Loud applause and amazed murmurs fill the air POV handheld shot from someone in the audience filming with a phone camera from the third row, frame slightly unsteady, raised above other people’s heads. The scene shows a large outdoor theater stage designed as an old European bakery. A baker stands defensively holding an oversized wooden rolling pin while facing a gigantic humanoid bread monster covered in crust and flour. The bread monster charges forward, shaking the stage with each step. The baker swings the rolling pin at the monster’s arms and torso while dodging massive sweeping attacks. Flour bursts into the air with every impact. Dramatic orchestral music plays through the stage speakers. The crowd can be heard cheering, laughing, and gasping as the baker narrowly avoids being crushed. The intense battle ends when the baker lands a powerful overhead strike to the monster’s head with the rolling pin, causing the bread monster to stagger, crack apart, and collapse heavily onto the stage floor. Audible oohs, applause, and excited cheering erupt once the battle is finished.

A hyper-real bodycam-style sequence in a dark, abandoned construction building at night. No ambient lighting—only helmet-mounted flashlights cutting through dust and darkness. Grainy image, slight motion blur, heavy breathing, radio static, muffled boots on concrete. Shot 1 (0–3s): The camera shakes as a squad kicks down a metal door. It crashes open. They enter a large, empty, decaying room—pillars, debris, broken walls. Flashlights sweep. At the far end— a man stands motionless, wearing a strange sunflower outfit, petals framing his head. He holds a lifeless woman in his arms. Shot 2 (3–6s): The team freezes. One shouts: “Drop her! Drop the woman—NOW!” Flashlights lock onto him. Dust floats in the beam. Silence. Shot 3 (6–8s): The figure suddenly lets out a distorted scream. In one motion—he throws the woman upward. His face splits open into a massive, unnatural mouth—elongated, alien. He snatches and consumes her mid-air. Shot 4 (8–12s): Chaos. The bodycam wearer stumbles backward, falls hard—camera tilts to the floor. From low angle— the creature launches forward, slamming into one marine—fast, violent, off-frame impact. It pivots instantly—takes down the second marine in front—screams cut abruptly. Shot 5 (12–15s): The creature turns toward the fallen camera. It screams directly into frame, petals shaking— Then leaps forward— Impact imminent— video cuts to black.

风格：8K 电影质感，真人实拍（非游戏感）。第一人称 POV——镜头即客人视角，轻微手持呼吸感，浅景深，她始终凑近镜头与"你"对视。暖色咖啡馆室内光，皮肤毛孔级真实、眼神湿润有眼神光，24fps，16:9。音频：环境咖啡馆白噪、她的台词与轻笑、衣物摩擦，无配乐无字幕。 SUBJECT —— @ （与输入图 100% 一致）：傲娇又温柔的可爱女仆，粉镜片圆框眼镜、棕色麻花辫，黑白女仆裙＋白围裙、白过膝袜、黑玛丽珍厚跟鞋。第一视角的女仆"情绪陪伴"——把全部注意力给镜头前的你，治愈、被宠、被关注的氛围。情绪节拍：温柔迎接 → 关心心疼 → 夸赞撒娇 → 宠溺定格。 LOCATION —— 温馨女仆咖啡馆角落卡座，暖黄灯、虚化吧台与花，STYLE REFERENCE ONLY，世界可自由延展。 ACTION —— 全程她对着镜头（你），眼神始终看进镜头，像你是这里唯一重要的人。 SHOT 1（0:00–0:04）—— 她在你对面坐下/微微俯身，双手托腮，眼睛弯弯看着你，软声："你回来啦～我等你好久了哦，主人。" SHOT 2（0:04–0:08）—— 神情转心疼，微微前倾凑近镜头，歪头打量你："今天……是不是又累坏了？没事的，在这儿你什么都不用想。" SHOT 3（0:08–0:12）—— 她弯眼笑起来，伸出一只手轻轻靠近镜头像要摸摸你的头/碰碰脸颊（手入框、不遮脸），宠溺地："你已经很棒很棒啦，我都看在眼里呢，乖。" SHOT 4（0:12–0:15）—— 傲娇地别过脸又忍不住转回来，脸颊泛红、轻声撒娇："就……只有今天，多陪你一会儿也可以的啦。" 定格在含笑望着你的眼睛上。 CAMERA —— 第一人称固定在客人坐姿高度、与她对坐，轻微手持呼吸晃动；全程平视她的脸，不切第三视角、不出现"你"自己。SHOT3 她伸手时镜头极轻微后仰，像被触碰。 STYLE —— 暖黄灯＋暖木卡座 60% / 黑白女仆裙 30% / 粉镜片＋脸颊红晕＋眼神光 10%。白平衡 4800K 暖调，柔焦温馨。 CONSTRAINTS —— 16:9，正常速度无慢动作。严格第一人称 POV，镜头即你的眼睛、不出现自己、不切旁观视角。她始终与镜头对视互动、传递被宠被治愈的情绪。手部入框真实受重力。健康温馨、无暧昧。面部发型服装 100% 一致。画面无标记符号。禁止眼睛发光。

[Global Settings] Format: 1:1 Style: Fast-paced fashion showcase, fisheye lens extreme wide-angle distortion, vibrant streetwear aesthetic, digital glitch transitions. References: Identity @Image1, Outfit references @Image2, @Image3, @Image4, Storyboard (layout / composition reference) @Image5. Audio/SFX: Upbeat fast-paced electronic track, digital glitch sound effects for transitions, city ambiance. Global Constraints: Character identity (face, hair, body type) strictly follows @Image1 across all beats. Storyboard @Image5 is for framing and shot order only. The protagonist is always interacting directly with the camera (first-person POV). [0:00-0:05] (Shot 1 & 2) Shot/Camera: First-person POV selfie perspective, fisheye lens extreme wide-angle distortion, dynamic handheld movement following the composition of frames 1-2 of @Image5. Subject/Action: Model wearing the FIRST reference outfit (cream fuzzy knit sweater, high-waisted light blue distressed straight-leg jeans, white low-top sneakers, brown leather crossbody bag, layered gold jewelry, white knit beanie). Model leans in closely to take a mirror selfie, waving playfully and moving dynamically close to the lens. Scene/Lighting: Busy city street intersection, vibrant color grading, bright daylight. Audio/SFX/Dialog: Fast electronic beat starts, street noise. Constraints: Outfit is fully visible and consistent, no missing items. Sharp focus on clothing details. Natural hands. [0:05-0:10] (Shot 3 & 4) Shot/Camera: Digital glitch transition into a continuous first-person POV fisheye lens, stepping back slightly, matching frames 3-4 of @Image5. Subject/Action: Model is now wearing the SECOND reference outfit (white cropped tee, high-waisted wide-leg gray trousers, brown leather loafers, brown leather belt, round sunglasses, small white crossbody bag). Model gestures confidently to the camera, then steps back to show the full outfit. Scene/Lighting: Minimalist urban sidewalk, crisp neutral color grading, soft natural shadows. Audio/SFX/Dialog: Sharp glitch sound effect at the start of the beat. Constraints: Exact outfit consistency, no replacement clothing, no outfit mixing. Identity remains locked to @Image1. [0:10-0:15] (Shot 5 & 6) Shot/Camera: Digital glitch transition into an extreme fisheye POV, dynamic handheld shake, following frames 5-6 of @Image5. Subject/Action: Model is now wearing the THIRD reference outfit (vintage blue denim jacket, white inner tee, high-waisted black straight-leg pants, black combat boots, white canvas crossbody bag). Model waves energetically and moves dynamically, leaning very close to the lens with an expressive attitude. Scene/Lighting: City street surrounded by towering skyscrapers, cinematic motion blur, hyper-realistic, vibrant tones. Audio/SFX/Dialog: Glitch sound effect, beat drop. Constraints: Consistent character identity. No off-model garments. Sharp focus on the denim and canvas textures despite the motion blur. [Negative Prompt] camera equipment, visible crew, microphones, production lights, extra characters, inconsistent faces, face or identity copied from the storyboard, identity drifting toward the storyboard render, wardrobe taken from the storyboard, inconsistent clothing, outfit mixing, wrong location, distorted hands, unreadable text, duplicate people, broken continuity, standard flat lens, missing accessories

FPV cinematic flight through a hyper-realistic treehouse at golden hour: bridges, staircases, warm interiors, exiting through skylight, fast-paced dynamic camera.

Immersive first-person FPV camera shot (camera never visible). The sequence begins outside the ancient Roman Colosseum at golden hour. The camera instantly accelerates at extreme speed toward the entrance and dives through the arches with explosive momentum. Inside, the corridor section is short and intense no wandering. The camera aggressively weaves through a brief stretch of curved Roman corridors with rapid snap turns, tight near-miss passes, and fast vertical shifts. It threads quickly past guards, torches, and stone archways, maintaining maximum velocity without slowing down. Within the first few seconds (well before the midpoint), the camera finds an arena access passage and immediately shoots upward. It bursts into the arena much earlier than before in one seamless motion. Now the focus shifts fully to the arena: a brutal survival fight with only 5–10 gladiators spread across the sand. The camera moves rapidly within the fight, not above it aggressively weaving between combatants. It flies directly through clashes: narrowly dodging swinging swords, smashing shields, stabbing spears, and lunging bodies. It curves tightly around individual duels, passes inches from impacts, slides past falling fighters, and constantly repositions at high speed. The movement feels reactive and alive quick direction changes, evasive maneuvers, and continuous motion through the chaos of the fight. Sand bursts into the air, bodies collide violently, and every pass feels dangerous and immediate. The camera spends most of the sequence inside the arena, dynamically navigating the fight from multiple angles, before finishing with a fast upward pull that briefly reveals the scale of the Colosseum. Hyper-realistic ancient Roman architecture, early arena reveal, ultra-fast FPV motion, aggressive weaving, dynamic exposure shifts (bright exterior → darker entry → harsh arena sunlight), realistic motion blur, cinematic lens effects, ultra-detailed textures, seamless one-shot, no cuts, no visible drone, high-intensity pacing, 4K, HDR.

Create one continuous 10-second ultra-photorealistic vertical 9:16 advertising video. REFERENCE IMAGES: @Image1 — MAIN CHARACTER + OUTFIT REFERENCE. Use ONLY the exact woman and exact fashion look from @Image1. Preserve her identity with extremely high consistency: the exact same facial structure, short glossy black bob haircut, side part, eyebrows, hazel-green eyes, nose, lips, smile, skin tone, body proportions and overall appearance. She must remain unmistakably the same woman throughout the entire video. Preserve the exact outfit from @Image1: oversized dark charcoal-gray patterned blazer, crisp white button-up shirt, blue-to-purple gradient necktie, matching charcoal-gray mini skirt, dark purple pointed slingback high heels, small round blue-purple lapel pin, purple ear accessory. Do NOT redesign, recolor, simplify or replace any part of the outfit. No wardrobe transformation. No hairstyle change. No face change. No additional jewelry. @Image2 — PRODUCT REFERENCE. Use the EXACT NEURA PEPTIDE 8 skincare jar from @Image2. Preserve its exact physical design throughout the entire sequence: frosted cool-white cylindrical glass jar, subtle lavender-purple gradient near the bottom, wide polished silver chrome lid, minimal premium luxury skincare aesthetic. Preserve the visible front branding: “NEURA” “PEPTIDE 8” “BARRIER REPAIR CREAM” The jar must always remain the same shape, proportions, material, lid size, colors and graphic layout. Never turn it into another cosmetic container. Never change it into a bottle, tube or different jar. LOCATION AND VISUAL STYLE: Downtown Manhattan, New York City. A huge luxury digital billboard mounted high on the façade of a modern building. The street below must feel ACTIVE and ALIVE throughout the entire video. Pedestrians are continuously walking naturally along the sidewalks and crossing the street. Different people move independently at realistic speeds. Some pedestrians pass in front of each other naturally. No frozen crowd. No static background people. Cars continuously move through the street during the entire shot: yellow taxis, sedans, SUVs, occasional delivery vehicle. Vehicles drive naturally through the intersection and along the road. Traffic must remain physically realistic and continuous. No frozen cars. No duplicated vehicles. No cars suddenly disappearing or teleporting. Real urban environment: Manhattan architecture, street traffic, yellow taxis, pedestrians, crosswalks, subtle scaffolding, street signs and realistic city details. Late-afternoon / soft golden-hour natural daylight. The video feels like a spontaneous real-life iPhone recording of an impossible luxury advertising installation. Handheld smartphone camera. Very subtle natural camera shake. Occasional tiny autofocus breathing. Realistic smartphone exposure. No artificial cinematic camera movements. No excessive motion blur. The billboard itself has a clean, premium beauty-campaign design: soft pearl-white / extremely pale cool-gray background, subtle silver reflections, very delicate blue-violet accents that visually complement the woman’s blue-purple tie and the lavender details of the NEURA jar. No additional advertising text on the billboard. No subtitles. No floating typography. SCENE 1 — BILLBOARD / 0.0–2.0 sec Start with a realistic street-level handheld iPhone shot looking slightly upward toward the enormous billboard. The entire billboard is clearly visible. The Manhattan street remains active in the lower part of the frame: people walk along the sidewalks, some pedestrians cross the street, yellow taxis and other cars drive naturally through the scene. Do NOT freeze the urban environment. Inside the billboard, the exact woman from @Image1 appears as part of a premium luxury skincare campaign. She is shown FULL BODY from head to toe. She is lying elegantly on her side across the lower part of the billboard composition: one arm supporting her upper body, torso slightly lifted, legs elegantly extended, purple slingback shoes completely visible. Her face remains clearly recognizable and photorealistic. She looks directly toward the real-world camera with a subtle confident smile. Her proportions must remain anatomically natural. The NEURA PEPTIDE 8 jar from @Image2 is displayed separately beside her in the right area of the billboard. IMPORTANT: the product must NOT overlap her head, body, arms, hands, legs, clothes or shoes. The product is large enough to be immediately recognizable but still realistically integrated into the billboard layout. The front of the jar faces the camera. “NEURA” and “PEPTIDE 8” remain clearly readable. At this stage both the woman and the product initially appear like realistic flat billboard advertising imagery. SCENE 2 — COME TO LIFE / 2.0–4.0 sec Without a cut, the woman gradually and realistically transitions from a flat 2D billboard advertisement into a real three-dimensional human being. The transformation must be subtle and physically believable. Depth develops naturally: her body gains realistic volume, her blazer fabric catches dimensional light, her hair moves slightly, her hands and legs become fully volumetric. Do NOT use glowing magic, particles, morphing effects or obvious AI transitions. She smoothly pushes herself upright using one hand and sits on the lower edge of the billboard. She crosses one leg over the other. Her upper leg begins gently swinging in a relaxed playful motion. Full body remains visible, including both purple shoes. Her exact face and short black bob remain unchanged. She maintains direct eye contact with the camera. The NEURA jar remains stationary in its original billboard position during this moment. Meanwhile, below the billboard: pedestrians continue walking, cars continue driving, traffic lights and street movement remain natural. The background city must remain alive during the entire transformation. SCENE 3 — TAKE THE PRODUCT / 4.0–5.5 sec Still sitting on the billboard edge, the woman notices the NEURA jar beside her. With a playful confident expression, she naturally reaches toward it with the hand closest to the product. Her hand physically wraps around the real jar. At the exact moment she lifts the product: the product disappears completely from its original printed billboard position. The area where the product was located becomes EMPTY clean billboard background. There must NOT be two copies of the product. The NEURA jar now exists exclusively in her hand as a real three-dimensional physical object. It has the exact same silhouette, frosted glass material, lavender gradient and silver chrome lid as @Image2. She briefly presents the front of the jar toward the camera. NEURA / PEPTIDE 8 stays facing the viewer. The Manhattan street below continues moving naturally. SCENE 4 — THROW TOWARD CAMERA / 5.5–7.5 sec The woman gives a subtle mischievous smile. She smoothly pulls her arm back and throws the NEURA jar directly OUT OF THE BILLBOARD toward the person filming. The jar physically exits the billboard space and enters the real Manhattan street environment. This is the main anamorphic illusion of the video. The jar rapidly travels toward the camera and becomes larger in perspective. It can rotate only VERY SLIGHTLY. Keep the front branding generally oriented toward the camera during flight. The product MUST NOT deform. NO: warped jar, melting glass, changing lid, missing lid, different packaging, extra products, duplicated product, changing colors, changing proportions, random text. The cylindrical jar remains rigid and physically correct. The camera operator instinctively reacts as the jar approaches: small realistic handheld jolt, tiny backward movement, brief natural autofocus adjustment. Pedestrians and moving cars remain visible in the background. The city never freezes. Do NOT cut away. SCENE 5 — FEMALE HAND CATCH / 7.5–9.0 sec A clearly FEMININE adult hand belonging to the woman filming enters naturally from the bottom foreground. IMPORTANT: the catching hand must unmistakably look female, NOT male. The hand should have: slender feminine fingers, delicate proportions, neat natural manicure, short-to-medium clean nails, elegant wrist, smooth feminine hand shape. No masculine hand. No large rough fingers. No hairy arm. No bulky wrist. No male-looking anatomy. The feminine hand catches the flying NEURA jar just before it reaches the phone camera. The catch is physically convincing: fingers wrap naturally around the jar, correct contact, correct perspective, correct scale. After catching it, the female hand holds the product very close to the lens. The jar dominates the foreground. Front label visible: NEURA PEPTIDE 8 BARRIER REPAIR CREAM The silver lid catches realistic Manhattan daylight reflections. Natural shallow smartphone focus: product sharp in foreground, billboard slightly softer in background. In the background, the woman remains seated on the billboard edge. She is still wearing the exact gray blazer, white shirt, blue-purple tie, gray mini skirt and purple shoes. She smiles toward the camera. Her legs remain crossed and the upper leg gently swings. The original product location on the billboard remains completely EMPTY. Pedestrians continue walking in the background. Cars continue moving naturally through the street. Yellow taxis pass through the scene. SCENE 6 — FINAL HOLD / 9.0–10.0 sec Hold this composition naturally for approximately one second. Foreground: the same feminine hand holds the NEURA PEPTIDE 8 jar close to the phone lens. The hand remains clearly feminine: slender fingers, neat manicure, delicate wrist. Background: huge Manhattan billboard with the same woman sitting on its edge. She makes subtle natural micro-movements: small smile, gentle breathing, tiny hair movement, soft leg swing. The urban environment remains continuously active: people walking, pedestrians crossing, yellow taxis and cars driving naturally. Nothing in the background freezes. Maintain the illusion that the skincare jar has physically traveled from the billboard into the real world. No cut. No transition. No freeze-frame. AUDIO: No music. No dialogue. No voice-over. Only realistic subtle ASMR environmental sound: NYC street ambience, moving traffic, distant taxi sounds, pedestrian ambience, soft city noise, subtle fabric movement, quiet hand/product interaction, a soft realistic air “whoosh” as the jar flies toward camera, a gentle physical catch sound. Do NOT make sounds exaggerated or cartoonish. ABSOLUTE CHARACTER LOCK: The woman MUST remain the exact woman from @Image1 throughout all 10 seconds. Exact face. Exact short black bob. Exact hair color. Exact body proportions. Exact outfit. Exact purple shoes. Exact tie. Exact blazer. Exact skirt. Exact lapel pin. No facial morphing. No beauty-filter face. No changing age. No changing ethnicity. No long hair. No hairstyle transformation. No duplicated character. No extra limbs. No malformed fingers. No disappearing shoes. ABSOLUTE PRODUCT LOCK: The NEURA PEPTIDE 8 jar must remain identical to @Image2. Same jar. Same silver lid. Same white frosted glass. Same lavender-purple lower gradient. Same proportions. Same branding placement. Never duplicate the product. Before she picks it up: ONE jar on billboard. After she picks it up: ZERO jar remaining in billboard graphic, ONE jar in her hand. During flight: ONE jar only. At final catch: ONE jar only in the FEMALE camera operator’s hand. The original billboard position stays permanently empty after the product is removed. PHYSICS: All body movements must follow realistic human anatomy. Natural sitting movement. Natural hand grip. Natural throw. Natural trajectory. Realistic gravity and inertia. Natural catch. The billboard structure must remain stable and architectural. The woman may sit naturally on its visual lower edge but never float. BACKGROUND ACTIVITY RULE: The NYC environment must stay alive during the entire 10 seconds. Pedestrians continuously walk. Some people cross the street. Vehicles continuously move through the road. Yellow taxis pass naturally. No frozen pedestrians. No frozen traffic. No repeated looping person. No duplicated cars. No teleportation. No disappearing vehicles. VISUAL QUALITY: Ultra-photorealistic. Premium global skincare advertising campaign. High-end fashion editorial aesthetics. Natural human skin texture. Realistic hair. Realistic fabric texture. Realistic polished chrome reflections. Realistic frosted cosmetic glass. Correct depth and perspective. Natural Manhattan daylight. The final result must feel like genuine iPhone footage of a spectacular real-world anamorphic billboard campaign, NOT like an AI-generated fantasy video. NO: cartoon look, CGI-looking skin, surreal deformation, beauty filter, face morphing, product distortion, duplicate jars, extra hands, extra fingers, male catching hand, floating objects, frozen pedestrians, frozen cars, unwanted text, subtitles, scene cuts, camera teleportation, music.

Remove the red lines, arrows, and all auxiliary markers. The camera is shot in a first-person (FPV) perspective, slowly, cinematically, in a single take, strictly following the start and end red lines in the image. Each flight lands on a different girl; the camera starts with the girl in yellow, then the girl in red, blue, green, and purple, before flying out of the window.

PART 1 (0:00–0:15) 0:00–0:01 Low-angle handheld selfie in front of Tokyo Tower, slightly tilted horizon, she adjusts the camera and smiles naturally. 0:01–0:02 Walking through Shibuya Crossing at night, handheld tracking shot, she turns back toward the camera while walking. 0:02–0:03 Inside a Japanese convenience store, she playfully holds an onigiri toward the camera and raises one eyebrow. 0:03–0:04 Sitting on a bench in Yoyogi Park, soft breeze moving her hair, she glances back toward the camera. 0:04–0:05 Close-up reaction while eating fresh takoyaki, laughing naturally after taking a bite. 0:05–0:06 Standing beside a glowing vending machine at night, casually pressing a button as colorful light washes across her face. 0:06–0:07 Reflection of her face in a moving train window at night, city light trails gliding across the glass. 0:07–0:08 Walking through a traditional torii gate, filmed from a low rear angle with natural walking motion. 0:08–0:09 Inside an immersive digital art exhibition, colorful lights sweeping across her face as she looks upward. 0:09–0:10 Low-angle wide shot beneath a giant Gundam statue, emphasizing its scale while she looks up excitedly. 0:10–0:12 Walking through a warm lantern-lit alley at night, filmed from behind, she briefly looks back over her shoulder with a smile. 0:12–0:15 Rooftop selfie overlooking the Tokyo skyline at blue hour, wind blowing strands of hair across the lens, glowing city lights behind her. PART 2 (0:15–0:30) 0:15–0:16 From the rooftop, she notices colorful amusement park lights in the distance and excitedly points the camera toward them. 0:16–0:18 Walking through Tokyo night streets toward the amusement park, handheld selfie footage with neon signs glowing around her. 0:18–0:19 Arriving at the amusement park entrance and looking up excitedly at a giant sky swing ride. 0:19–0:20 Buying a ticket at the ticket counter. 0:20–0:21 Holding the ticket toward the camera while entering the amusement park. 0:21–0:22 Waiting in line, smiling nervously and looking up at the ride. 0:22–0:23 Sitting in the giant swing seat as the safety bar locks into place. 0:23–0:24 The ride begins rising slowly into the night sky, anticipation building on her face. 0:24–0:25 The swing climbs higher, wind blowing through her hair as she starts laughing. 0:25–0:26 The ride accelerates dramatically, she screams and laughs with excitement. 0:26–0:27 Full-speed swing motion, dynamic POV shots, city lights streaking behind her. 0:27–0:28 Highest point POV, breathtaking Tokyo skyline and city lights below. 0:28–0:29 The ride slows down, she catches her breath and laughs. 0:29–0:30 Walking away through the illuminated amusement park, turning back with a huge smile and thumbs-up as the vlog fades out.

"A single continuous handheld handycam shot in slow motion, drifting and weaving organically through a completely frozen party scene, every person suspended mid-movement like a living diorama. The shaky camera circles and floats slowly. Single seamless slow-motion take."

Format: Found-footage vacation video, 15 seconds, seven shots with hard cuts — a tourist's camera roll from one full day, morning to dusk. Image reference: Use the attached gate photo as the exact visual reference for Shot 1 — match its architecture, proportions, torch placement, lighting, and signage precisely. Main subject: Young woman, blond hair in a loose low bun shedding strands, oversized faded olive t-shirt, denim shorts, white canvas sneakers, disposable-camera strap on her wrist. Realistic skin texture, minimal makeup, wide-eyed tourist joy. Maintain perfect visual consistency — identity, hair, clothing, proportions — across every shot and location. Location: A vast island dinosaur preserve. Jungle mountains, open savanna paddocks, wooden observation structures, tall electrified perimeter fences, dirt tour roads. Dinosaurs are photorealistic animals — muted natural hides, birdlike motion, visible breathing weight. Filmed like wildlife, never like monsters. Camera style: Mid-range smartphone, handheld by her friend. Heavy shake, autofocus hunting, motion blur on pans, compression grain, exposure pumping. Each cut is abrupt and mid-motion, like real unedited camera-roll clips. 00:00–00:03 — THE GATES (morning, filmed from inside a moving open-top tour jeep). The camera films forward over the seats as the jeep approaches the colossal timber gates from the reference image — flared stone-and-wood towers, flaming torches, jungle pressing in on both sides, the road still wet from morning rain. The doors groan open. She spins around from the front seat into frame, gripping the roll bar, and half-shouts over the engine: "Oh my god, this is wild! I can't believe we're actually here!" — as the jeep passes under the archway into shadow and out again. 00:03–00:05 — SAVANNA (late morning, wide shot from a raised wooden platform). She stands small at the rail against an open plain where a herd of long-necked giants wades through heat haze, one rearing slowly against the treeline. The camera zooms in shakily, overshoots, corrects. 00:05–00:07 — RIVER CROSSING (midday, low angle through the jeep's open side). Past her shoulder as the jeep fords a shallow river: a crested duck-billed dinosaur drinks twenty meters away, lifting its head to watch them pass, water dripping from its jaw. She raises her disposable camera; the phone catches its tiny click. 00:07–00:09 — JUNGLE FENCE (afternoon, handheld walking shot). Dim under the canopy beside a tall humming fence. She peers through the wire — and a small crested predator peers back from the ferns, head tilting in quick birdlike jerks. She backs up a slow half-step, whispering "okay... hi." Autofocus argues between wire and eye. 00:09–00:11 — THE HERD RUNS (late afternoon, chaotic pan from the jeep at speed). A flock of small striped dinosaurs floods across and around the vehicle. The camera whips left-right, blurred bodies steaking past, her whooping off-mic, dust on the lens. 00:11–00:13 — QUIET GIANT (golden hour, static shot at a wooden overlook). The frame finally settles: silhouetted against low sun, she stands at a rail as an immense horned dinosaur grazes just below, close enough to hear grass tearing. She isn't laughing anymore — just watching. 00:13–00:15 — DUSK ROAR (blue hour, half-framed and accidental). Filmed as the phone was being lowered, frame tilted: a colossal, deep bellow rolls in from the darkening jungle far off. Everyone freezes and turns toward it. Cut to black mid-turn. Audio: Ambient only, shifting per location — jeep engine and gate groan, her excited line over the motor, savanna wind, river wading, fence hum, thundering small feet, grass tearing, and the single distant bellow that swallows everything. No music. No narration. Style & quality boosters: Raw consumer camera-roll aesthetic, abrupt mid-motion cuts, heavy handheld instability, natural motion blur, golden-hour lens flare, coherent animal mass and physics — ground tremor, water displacement, dust. No text overlays, no watermarks. Goal: One unforgettable day inside the world's most famous dinosaur preserve, told through seven imperfect tourist clips — opening on the gates everyone recognizes, and closing on the sound that reminds them what lives beyond the fences.

Style: Ultra realistic AAA open-world gameplay. Inspired by modern crime-action sandbox games. Third-person gameplay camera with occasional first-person interaction shots. Natural physics, cinematic lighting, dynamic reflections, ray-traced puddles, realistic water simulation, 60 FPS gameplay feel, authentic HUD, no cinematic movie shots. Everything should feel like a real gameplay recording from a leaked next-generation game. Video Length: 25-35 seconds Scene 1 (Mission Start) Player controls Lucia walking toward a colorful car wash. Camera follows behind her with slight handheld gameplay movement. A mission notification smoothly appears: NEW SIDE MISSION CAR WASH "Clean the customer's vehicle." Reward: $600 + Respect Mini map, money counter, objective tracker, interaction prompt and subtle gameplay HUD are visible. Audio: Open-world city ambience, distant traffic, birds, passing cars, pedestrians talking, footsteps on wet concrete, mission notification sound. Music: Low-volume GTA-style upbeat electronic background music. Scene 2 (Pressure Wash) Lucia grabs the pressure washer. Player aims freely around the vehicle. Strong water pressure blasts away dirt. Water splashes naturally. Soap begins forming. Camera has realistic player aiming movement with tiny imperfections. Audio: High-pressure water spray Water hitting metal Splashing puddles Wet concrete sounds Trigger click Controller vibration feel Scene 3 (Soap Application) Player switches tools. Close gameplay interaction of sponge moving across hood, windows and doors. Soap foam reacts naturally. Circular hand movements. Foam drips downward. Audio: Sponge rubbing paint Foam squishing Water dripping Hand movement sounds Scene 4 (Rinse) Player sprays entire vehicle again. Soap washes away dynamically. Water runs off mirrors. Reflections become cleaner. Sunlight sparkles on wet paint. Audio: Continuous water spray Heavy splashes Water flowing Drain sounds Scene 5 (Wheel Cleaning) Gameplay zooms toward wheels. Player scrubs rims. Then applies tire shine spray. The tire instantly becomes glossy. Audio: Brush scrubbing rubber Spray can hiss Metal wheel sounds Tiny droplets Scene 6 (Drying) Player wipes remaining water using microfiber towel. Camera follows hand naturally. Paint becomes mirror-like. Audio: Microfiber rubbing Small squeaks Soft cloth movement Scene 7 (Inspection) Player slowly walks around the finished vehicle. Camera circles naturally with gameplay joystick movement. Objective updates: Vehicle Cleaned 100% Audio: Footsteps Ambient city Small UI completion sound Scene 8 (Mission Complete) Customer walks over smiling. Hands Lucia cash. Mission overlay appears exactly like gameplay: MISSION PASSED CAR WASH +$600 +250 RP Cash counter increases. Audio: Mission complete jingle Cash register sound NPC: "Looks brand new! Thanks!" Scene 9 (Drive Away) Lucia enters the cleaned vehicle. Engine starts. Reverse lights activate. She slowly exits the car wash into busy city streets. Gameplay camera remains behind the car. Audio: Door open Door close Seat movement Engine ignition Idle exhaust Gear shift Tires rolling over wet pavement Traffic ambience Background Music: Modern upbeat electronic gameplay soundtrack. Subtle, never overpowering sound effects. Gameplay Camera Rules: • Slight controller shake • Small joystick corrections • Natural acceleration/deceleration • Smooth third-person gameplay camera • Authentic HUD always visible • No cinematic cuts • Looks like genuine leaked gameplay footage Quality: 4K photorealistic, ultra detailed, realistic water physics, dynamic reflections, volumetric lighting, motion blur, ambient occlusion, physically accurate materials, next-generation graphics, highly immersive gameplay recording.

Create a realistic cinematic 15-second vertical travel vlog video in 9:16 format. A beautiful Western woman tourist is walking through the lively streets of Yogyakarta, Indonesia, while recording herself with a smartphone in selfie vlog style. She looks cheerful, curious, and friendly. The background shows an authentic Jogja street atmosphere with local food stalls, motorcycles passing slowly, pedestrians, Indonesian street signs, warm afternoon sunlight, and a cozy street food vibe. Scene breakdown: 0–3 seconds: The woman walks along a lively street in Yogyakarta while holding her phone camera. She smiles at the camera and speaks naturally in English: “Hey guys, I’m here in Yogyakarta, Indonesia, and today I’m trying some local street food!” Audio: natural street ambience, soft motorcycle sounds, people chatting in the background, street food sizzling faintly. 3–6 seconds: She approaches a small street food stall selling bakso bakar. The vendor grills meatball skewers over charcoal. Smoke rises beautifully, and the sauce is brushed onto the bakso. The woman looks excited and points the camera toward the food. She says: “Look at this! Grilled bakso… it smells amazing!” Audio: sizzling grill sound, light charcoal crackle, vendor activity, street ambience. 6–10 seconds: Close-up shot of the woman holding a skewer of bakso bakar. She blows on it gently, takes a bite, then pauses for a moment. Her eyes widen, she smiles happily, and gives a delicious-food expression. She reacts: “Mmm! Wow… that is so good!” Expression: happy surprise, eyes slightly widened, smiling while chewing naturally, enjoying the taste. Audio: soft bite sound, subtle “mmm” reaction, background grill sizzling. 10–15 seconds: She faces the camera while holding the bakso bakar close to the lens. She reviews the taste enthusiastically but naturally. She says: “It’s smoky, sweet, spicy, and super savory. This is seriously delicious. You have to try this in Jogja!” Final shot: she smiles warmly, gives a thumbs up, and the bakso bakar stall remains visible in the background. Visual style: ultra realistic, cinematic travel vlog, natural handheld camera movement, slight vlog-style shake, warm natural lighting, vibrant but realistic colors, authentic Indonesian street food atmosphere, shallow depth of field, detailed face, realistic body movement, natural facial expressions, 4K quality. Character details: beautiful Western woman, around 25 years old, casual summer outfit, natural makeup, friendly travel vlogger personality, expressive and confident. Audio style: realistic English female voice, cheerful travel vlog tone, natural street ambience, grill sizzling, light traffic, people talking softly in the background, no music overpowering the voice. Negative prompt: blurry face, distorted hands, extra fingers, unnatural mouth movement, robotic voice, bad lip sync, fake expression, overacting, low quality, incorrect Indonesian setting, empty street, unrealistic food, text glitches, watermark, logo, bad anatomy, overexposed lighting.

Seedance2.0 - Beat Synced Scrambled Routine FORMAT: 15s / 145 BPM / 15 SHOTS / beat-synced scramble routine SUBJECT: Battle-scarred orbital pilot with grease-stained skin. WARDROBE: Soiled thermal undergarments inside. Heavy, scratched pressure suit and shattered visor helmet outside. ENVIRONMENT: Cramped zero-G berth, flickering red emergency strobes, rusted airlock, cramped cockpit, stark black void of space with laser fire. MOOD: Adrenaline shock, desperate scramble, high-G physical strain, exhausted deflation. MUSIC: Aggressive industrial synth-metal. COLOR LOGIC: Gritty Cyber-Industrial Look. STYLE: Ultra-Realistic. LOGIC RULE: Keep logical consistency in wardrobe, props, locations, and action continuity across all shots. SHOT 1: ECU, 85mm push-in / Flashing red warning light reflecting in a dilated pupil as the eye snaps open. / SFX: klaxon blare, heavy inhale. SHOT 2: WS, 35mm handheld jolt / Rhythmic cut into the pilot kicking off the metal bulkhead, floating violently across the zero-G berth, thermal undergarments clinging to sweat. / SFX: metal clang, fabric rip, fast breath. SHOT 3: MCU, 50mm slide / Cut on action into a magnetic boot locking onto a steel deck plate, sparks kicking up from the impact. / SFX: heavy mechanical clack, spark hiss. SHOT 4: Insert shot, 85mm rack focus / Match cut into pressure gloves snapping onto wrist seals, twisting violently to lock, green indicator light flaring. / SFX: pneumatic hiss, plastic click. SHOT 5: Interior locker view, 24mm wide / Object pass into the camera inside the gear locker looking out as the heavy helmet is ripped from the rack, red strobe light illuminating a panicked grimace. / SFX: metal rattle, heavy scrape. SHOT 6: Insert shot, 50mm handheld / Rhythmic cut into the helmet slamming over the head, latches snapping shut beneath the chin. / SFX: heavy thud, air pressurization whine. SHOT 7: MCU, centered 50mm push-in / Match cut into a violent thrust forward, hand slamming a yellow airlock release button. / SFX: solid smack, pneumatic release. SHOT 8: Bird's-eye insert, 35mm overhead / Cut on action into heavy thruster pack igniting, blue flame violently scorching the deck. / SFX: thruster roar, metal groan. SHOT 9: MS, 35mm pivot / Camera wipe into a chaotic trajectory through the airlock, thermal suit completely covered by the scratched pressure suit, tearing through floating debris. / SFX: wind rush, debris clatter. SHOT 10: Insert shot, 50mm overhead / Match cut into a thick throttle lever being yanked entirely backward in one aggressive pull. / SFX: mechanical clank, heavy friction. SHOT 11: WS, 24mm parallax / Whip pan transition into the massive rusted mech tearing out of the hangar bay into the black void of space, thrusters leaving a hard white trail. / SFX: vacuum silence, low bass rumble. SHOT 12: MS to CU, 35mm glide into 85mm push-in / Sound bridge into the cockpit interior as the pilot grips the dual joysticks, body vibrating with G-force, eyes darting across radar pings, dodging a blinding green laser blast that chars the hull outside the glass. / SFX: radar ping, hull creak, heavy breathing, laser sizzle. SHOT 13: Insert to MCU, 50mm snap zoom / Smash cut to the primary weapon trigger as a thumb crushes the red button, recoil shaking the frame violently. / SFX: cannon boom, metallic recoil. SHOT 14: OTS, 35mm handheld / Rhythmic cut into targeting reticles locking, warning screens flashing yellow, sweat violently flinging from the pilot's brow as the cockpit violently rolls. / SFX: warning buzzer, violent thruster hiss. SHOT 15: WS, 50mm pull-out / L-cut with a match from the targeting screen close to the berth re-entry as the pressure suit drops, helmet rolls across the deck, and the pilot floats backward in thermal wear, collapsing into the netting in the opening frame shape. / SFX: airlock hiss, heavy thud, deep exhale, room tone.

参照画像をファーストフレームとして使用する。 15秒、24fps。 日本語アイドルソング付きの短尺MV。 映像全体を通して、ファーストフレームの超広角・魚眼セルフィーレンズ構図を維持する。 通常レンズに戻さない。 カメラを切り替えない。 画角を補正しない。 魚眼レンズ特有の歪み、近接パース、顔や手が大きく見える遠近感を最後まで維持する。 カメラは超広角魚眼レンズを使用した固定カメラ。 映像全体を通してカメラ位置は維持する。 キャラクターのみがレンズへ近付いたり離れたりする。 レンズの歪み、近接パース、遠近感は最後まで維持する。 観客に向けたステージダンスではなく、レンズ一台に向かって全力でアイドルパフォーマンスをするセルフィーMV。 キャラクターの顔、髪型、髪色、瞳、衣装、体型、雰囲気、デザインはファーストフレームから絶対に変えない。 キャラクターを別人化しない。 衣装を変更しない。 映像スタイル： 高品質な日本アニメMV。 滑らかな作画。 表情豊か。 魚眼レンズの迫力を活かした近距離パフォーマンス。 音楽： 明るくキャッチーな日本語アイドルソング。 BPM高め。 可愛い女性ボーカル。 短く覚えやすいサビ。 電子音を含むポップなアイドル楽曲。 歌詞： 「ねぇ！ こっち見て！ 今日はキミだけ 特等席 ドキドキ 止まらない だ・い・す・き！」 タイムライン： 【0.0〜1.5秒】 ファーストフレームと完全一致して開始。 キャラクターがレンズを覗き込み、少しだけ目を近づける。 魚眼で瞳が大きく迫って見える。 歌詞： 「ねぇ！」 【1.5〜3.0秒】 キャラクターがカメラに向かって人差し指を差す。 指が魚眼レンズで大きく誇張される。 小さく笑う。 歌詞： 「こっち見て！」 【3.0〜5.0秒】 リズムに合わせて左右に揺れる。 歌詞： 「今日はキミだけ」 【5.0〜7.0秒】 キャラクターがレンズにさらに近づく。 片目が画面いっぱいに近づき、ウインクする。 可愛く挑発するような表情。 歌詞： 「特等席」 【7.0〜9.0秒】 両手でレンズの周囲を囲み、ファーストフレームのフレームポーズを変形させる。 指が手前で大きく見える。 歌詞： 「ドキドキ」 【9.0〜11.0秒】 レンズを軽くトントンと叩く。 画面が小さく揺れる。 レンズ接触による一瞬の軽いブレを入れる。 歌詞： 「止まらない」 【11.0〜13.5秒】 キャラクターが笑顔で顔を近づけ、カメラ目線で小さく踊る。 肩、頭、手だけを使った近距離アイドルダンス。 全身ダンスにはしない。 歌詞： 「だ・い・す・き！」 【13.5〜15.0秒】 最後にキャラクターが手のひらでレンズを覆う。 画面が魚眼のまま暗くなり、ブラックアウトして終了。 禁止事項： カメラ切り替え禁止。 通常レンズ化禁止。 魚眼歪みの補正禁止。 遠景の全身ダンス禁止。 ステージ俯瞰禁止。 キャラクターの顔や衣装の変更禁止。 別人化禁止。 実写化禁止。 背景の大幅変更禁止。 歌詞を英語にしない。 字幕を表示しない。

Create a 15-second, vertical 9:16 ultra-realistic video filmed in ONE continuous handheld shot by an unseen friend on a high coastal cliff above the sea. A woman makes a running jump from the cliff onto an enormous floating trampoline far below, rebounds upward, performs one backward somersault, and lands back on the SAME cliff before celebrating. The image feels like genuine smartphone footage: ordinary hand shake, slightly imperfect framing, natural motion blur, mild compression, and small autofocus and exposure adjustments. One viewpoint and one wide-angle lens throughout. MAIN SUBJECT An exceptionally beautiful 24-year-old platinum-blonde woman with a striking feminine face, expressive eyes, full lips, a narrow waist, a very curvy hourglass figure. Overall she is slim, but have a very large full bust, wide rounded hips, and very large, full, rounded buttocks. Her long hair is tied in a high ponytail. She wears a vivid hot-pink athletic bikini: a fitted supportive top with a low neckline and a secure underband, plus matching high-cut thong bottoms with narrow side bands. The thong has a narrow rear strip, leaving her rounded buttocks uncovered. The fabric is opaque and remains securely in place. White athletic sneakers. Natural skin texture, light makeup, realistic fabric tension, and believable movement. Preserve her face, figure, ponytail, bikini, and footwear from beginning to end. CLIFF AND SPATIAL CONTINUITY A broad, pale limestone clifftop overlooks a deep turquoise sea. The sheer rock face drops dramatically beneath the takeoff edge. Small waves and white foam around the cliff’s base establish the enormous height. The clifftop has a flat, clear running path and a broad landing area beside the takeoff point. Distinctive cracks in the pale rock remain recognizable throughout. Four friends stand together to the LEFT of her running path. They watch and record her with their phones, keeping both the takeoff and return landing areas clear. Far below, just seaward of the cliff face, is an ENORMOUS circular floating water trampoline, many times wider than a person. It has a thick white inflatable flotation ring, bright cobalt-blue padding around its inner edge, and a broad dark-grey elastic jumping mat stretched across the centre. The mat is visibly elevated above the seawater. The trampoline is already positioned below her outward jumping path from the beginning. It remains moored in the SAME location relative to the cliff. Small waves move around its floating ring. The trampoline’s centre mat, rather than its inflatable rim, receives her landing. CAMERA The unseen friend films from behind and slightly to the side of her final running steps, then remains on the clifftop beside the takeoff area with an unobstructed view down toward the trampoline. The viewpoint continuously follows her jump downward, then follows her rebound upward and back onto the same cliff. The camera operator stays on the clifftop throughout. Keep her visible through the descent, mat contact, rebound, and somersault. She naturally becomes smaller as she falls away and larger as she returns. Use the pale cliff edge, the vertical rock face, the trampoline’s white ring, and the sea as consistent spatial landmarks. The recording device and its operator remain outside the image. Her friends’ phones may be visible in their hands. CONTINUOUS ACTION 00:00–00:02 — IMMEDIATE JUMP Start with her ALREADY running toward the cliff edge. Her final strides, the open sea beyond her, and the friends watching from the left are immediately visible. She takes her last few running steps and jumps outward within the first TWO seconds. Her feet visibly leave the rock. The opening moves directly into the jump without a countdown or hesitation. 00:02–00:05 — THE DROP AND CONTACT The view follows her over the edge and tilts downward in the same uninterrupted movement. Reveal the enormous floating trampoline beneath her, surrounded by seawater. The sheer cliff face and part of the same clifftop edge remain visible along the side of the image, making the height and filming position clear. She descends feet first toward the CENTRE of the dark jumping mat. Show ONE clear contact: both shoes meet the mat, her knees bend, and the elastic surface depresses into a deep bowl beneath her weight. The deformation spreads outward toward the padded circular edge. The white flotation ring briefly rocks under the load while remaining in its established location. 00:05–00:10 — REBOUND AND RETURN The depressed mat springs back, launching her upward and slightly back toward the SAME clifftop. Her feet visibly separate from the mat before she begins rotating. The trampoline stays on the sea below; its mat continues oscillating briefly after her departure. The camera follows her ascent without losing her or changing viewpoint. During the rising return arc, she performs ONE spectacular tucked backward somersault. Show one continuous, readable rotation: she tucks, rotates backward once, then opens her body and extends her legs toward the landing area. Her movement carries her back toward the original takeoff ledge, with clearance from the vertical rock face. She clears the cliff lip and lands with BOTH feet on the flat rock beside her takeoff point, approximately nine to ten seconds into the recording. Her knees bend to absorb the landing. Her shoes visibly contact the limestone, and her ponytail continues moving before settling. 00:10–00:15 — THE CELEBRATION Only after she has landed and regained her balance does she turn toward the nearby camera. Her friends erupt into cheers, jumping and raising their free hands while leaving her route clear. She throws her arms up and shouts, breathless and ecstatic: “YES! I DID IT!” She runs the short distance toward the camera, laughing with relief and excitement. The person filming reacts with a small natural shake and a couple of backward steps farther inland as she approaches. Finish close to her radiant, laughing face as her friends gather beside and behind her, cheering. The coastal setting remains recognizable around them. End naturally at 15 seconds while the celebration is still happening. LIGHTING AND SOUND Bright natural coastal daylight, with thin cloud softening the sunlight. Readable pale rock, deep blue-green water, natural skin tones, and vivid pink fabric. Small sea reflections may briefly clip in exposure. The camera adjusts slightly when tilting between the bright water and her face. Preserve ordinary phone-camera contrast and skin texture. Natural recorded sound only: shoes striking rock, coastal wind, distant waves, a deep elastic thump at trampoline contact, the solid clifftop landing, cheering, laughter, and her short English shout. Voices have natural breaths and phone-microphone character. No music, narration, or cartoon sound effects. ESSENTIAL CONTINUITY One running jump → one visible feet-first contact with the trampoline’s central mat → one rebound → one backward somersault → one landing on the original clifftop → celebration. One continuous recording at normal speed. No cuts, hidden transitions, slow motion, external aerial views, or separate reaction shots. The return is a new upward movement following the visible rebound, never reversed falling footage. The trampoline remains below on the sea. Its elastic mat compresses and rebounds; the whole trampoline does not rise into the air with her. She returns to the SAME clifftop beside the SAME friends. Preserve the cliff geometry, sea level, trampoline position, daylight, and her appearance. No hovering, teleportation, duplicated people, changing clothes, water landing, captions, subtitles, logos, watermarks, or recording interface.

In 2026, create a 15-second cinematic travel video featuring characters who travel to 30 world-famous locations. The frame rate is synchronized with the rhythm every 0.5 seconds. Portable camera on a selfie stick, wide-angle lens, clear selfie framing, energetic travel video blogger style, bright cinematic colors, realistic lighting, dynamic motion blur, natural crowd and clear landmarks in each shot. a cheerful face changes, gender changes, hairstyle changes, wristwatch changes, or clothing changes. Venues: Eiffel Tower in Paris, Shibuya in Tokyo, Times Square in New York, Colosseum in Rome, Pyramids in Egypt, Cathedral of Christ the Savior in Rio de Janeiro, Big Ben in London, Sydney Opera House, Grand Palace in Bangkok, Taj Mahal in India, The Great Wall of China in China, Burj Khalifa in Dubai, St.Sophia Cathedral in Istanbul, Grand Canal in Venice, Machu Picchu in Peru, Acropolis of Athens, Sagrada Familia Cathedral in Barcelona, Amsterdam Windmills, Gyeongbokgung Strict identity in all scenes: the same face, the same age, the same hairstyle, the same body proportions, the same palace in Seoul, Marina Bay Sands in Singapore, Santorini in Greece, Petra in Jordan, Neuschwanstein Castle in Germany, the Golden Gate Bridge in San FranciscoSan Francisco, Niagara Falls in Canada, Mount Fuji in Japan, Angkor Wat in Cambodia, the Swiss Alps in Switzerland, Turkish Cappadocia and the Maldives.Each place features a unique stylish outfit that matches the local atmosphere, and funny expressive poses such as waving arms, peace signs, namaste, finger hearts, laughter, pointing out landmarks, thumbs up, surprised reactions, dancing or playful gestures while sightseeing.

Bir kartalın POV bakış açısından sokakta farenin kovalanması, tek plan çekim. bazı insanlar fareyi görünce kaçıyor. Fare ve kartal ayaktaki ve kafede oturan insanların ayaklarının arasından geçiyorlar, hızlı ve dinamik bir kovalamaca videosu olmalı. ultra realistik ve sinematik. Belgesel sahnesi gerçekçiliğinde

First-person FPV drone shot with cinematic. Motion pattern: Fly-in >close up >greeting >next character