AI Prompt Library

Create a 15-second high-fashion football shoe advertisement using the uploaded reference images for the gold metallic football cleats, white knee-high socks, and vintage feminine football styling. Keep the footwear design, colors, textures, and proportions consistent throughout the entire video. Environment: An enormous open outdoor football field, covered in lush natural green grass, with a completely open horizon and a beautiful dreamy pastel blue-and-pink sky. The field should feel spacious and untouched. No stadium, no spectators, no buildings, no bleachers, no indoor environment, no close fences. 0–3 sec: Wide cinematic shot of a young adult female footballer standing alone in the middle of the vast open field. She wears the same elegant vintage-inspired outfit from the reference images with white knee-high socks and the gold metallic football cleats. She gently places one foot on a football. Camera slowly moves toward her while sunlight reflects beautifully off the cleats. 3–6 sec: Cut to dramatic close-up shots of the gold cleats. Her foot steps firmly onto the grass, the studs press naturally into the turf, and she tightens the laces. Macro camera movement reveals the metallic texture, stitching, sole, and studs. Make the shoes the visual hero. 6–10 sec: She starts playing football across the completely open field, dribbling the ball quickly and confidently. Use a low-angle tracking camera focused on her feet and cleats as she changes direction, accelerates, and controls the ball. Her flowing outfit moves naturally with her movement. 10–13 sec: She takes a powerful run-up and strikes the football with the gold cleat. Capture the kick in cinematic slow motion, showing the shoe making contact with the ball, grass particles flying naturally, and the ball launching across the open field. 13–15 sec: Premium hero shot. The football rolls to a stop beside her gold cleats. She stands confidently in the background on the vast green field as warm sunlight catches the metallic shoes. Camera slowly pushes toward the cleats and ends on a sharp product-focused close-up. Visual style: luxury sports fashion campaign, dreamy editorial photography, cinematic natural lighting, realistic football movement, realistic grass physics, elegant feminine styling, dynamic camera movement, macro shoe details, shallow depth of field, premium commercial quality, photorealistic, subtle film texture. Important: Keep the gold football cleats identical throughout the video. No changing shoe design, no distorted feet, no extra limbs, no warped football, no artificial CGI appearance. The entire video takes place outdoors on one vast open football field.

Create a cinematic, ultra-realistic 9:16 school classroom scene. A young male student with styled red hair suddenly reacts dramatically as a large burst of flames erupts from a backpack in front of him. Another student in a school uniform sits at a desk in the background, shocked and watching. Warm classroom lighting, realistic fire and smoke effects, dynamic movement, natural facial expressions, detailed environment, cinematic camera motion, shallow depth of field, realistic textures, dramatic atmosphere, high detail, 4K, photorealistic.

Cinematic realistic texture, modern Chinese workplace drama. Centered on sharp but restrained social observation, colloquial language easily understood by ordinary people, senior screenwriter-style dialogue close to daily communication habits, and subtle, authentic character performances. Adopt: Arri Alexa film camera texture Stable and clear facial micro-details Slight film grain Realistic lighting formed by mixing office ceiling lights and natural light from outside the window Core Theme Directly recreate the core situation of the current meme "Take 3 days off, rest for 13" without any ancient style translation. What needs to be performed is not "teaching everyone how to combine holidays," nor the cliché "bad boss refuses leave - employee boldly slaps back." Dig one layer deeper into reality: Many people are clearly using their own vacation time, yet they have already developed a habit of proving themselves before speaking up: "I'm not lazy" Won't drag others down Work is arranged I am still responsible enough Even feeling guilty proactively at the end. The real twist isn't the senior sister winning an argument for the junior sister, but suddenly turning this matter, which everyone has made so serious, back into a normal small thing: Handover done, go rest. Character & Environment Anchors Strictly use @Image 1 and @Image 2 as facial identity anchors for the characters, while naturally adjusting clothing according to the modern workplace plot. All uploaded background and location references should first be silently reorganized into a unified credible modern space, including: Real architectural scale Office desks and chairs Glass reflections Walking colleagues Computer screens Elevator dynamics Main light direction from outside the window Air depth Environment remains vivid but must never actively create plot. Role Settings Role ID A | Project Lead Always the same woman from @Image 1: 27–32 year old East Asian female Oval fair face Dark almond eyes Black long hair half-up Tall and slender figure Identity in this scene: Steady project lead Modern clothing Ivory white shirt Dark grey slim trousers Black leather shoes Simple silver wristwatch Role ID B | Junior Employee Always the same woman from @Image 2: 22–26 year old East Asian female Round lively face Black braided hair Petite figure Identity in this scene: Junior employee Modern clothing Green knit top Loose dark trousers White sneakers Always holding the same filled leave application form. Minor Roles One department manager Two colleagues Exist only as minor witnesses. Segment Structure 0–5s | Wide shot gradually pushing in slightly Manager looks at the leave form, says half-jokingly and half-pressure: "Three days off? Rest thirteen? You really dare." Junior sister character immediately starts explaining: "I still have three days annual leave..." Manager continues: "If everyone takes leave like you, who does the work?" At this moment, the confidence the junior sister had prepared visibly collapses. Do not show anger, only present that hesitation familiar to many ordinary office workers: "Suddenly feeling like I did something wrong" 5–10s | Medium two-shot relationship lens + extremely short reaction close-ups Maintain: Same characters Same clothing Same leave form Completely consistent spatial relationships Junior sister slowly reaches out to take back the leave form, whispers: "Maybe... I won't ask for it." The senior sister character, who hasn't spoken all along, finally opens her mouth: "Wait." She doesn't argue with the manager first, but directly asks the junior sister: "Is the handover done?" Junior sister: "Done." Senior sister: "Did you leave emergency contacts?" Junior sister: "Yes." Senior sister takes the leave form, signs it very naturally, hands it back to her, and says only: "Then go rest." The entire action must be calm to the point of being ordinary. Let the audience feel the first contrast: Everyone else made this matter so serious, but she just treats it as a normal workflow. 10–15s | Close-up entering extreme close-up Manager finally frowns: "You approved it just like that?" Senior sister looks at him: "Otherwise what?" Complete pause for half a beat. Junior sister holds the leave form, still retaining some guilt, lowers voice and asks: "Will it look like I'm irresponsible?" Senior sister truly looks at her this time: "Responsibility means handing things over clearly." Pause for half a beat. No chicken soup, no slogans, use the most ordinary human words to land the final sentence: "Not taking a vacation, and feeling like you owe someone." Junior sister finally exhales the breath she held, puts the leave form in her bag. Manager is not forcibly slapped or suddenly admits fault, just momentarily speechless. Shallow depth of field background: One colleague silently closes the work group chat on phone Another colleague suppresses a light knowing smile Finally: No cheering No applause No victory music Junior sister walks to the elevator herself. Senior sister has naturally turned back to continue working. The whole event ultimately becomes something that was supposed to be very ordinary. Performance Principles Character performances must be realistic and restrained. Junior Sister Focus on expressing: Fully prepared at start Confidence crushed by one sentence from manager Instinctive guilt arises Wants to withdraw her normal need proactively Finally truly relaxes No anger, no crying, no short-drama style eye-bulging. Senior Sister Focus on expressing: Observing throughout first half Doesn't rush to fight for junior sister First confirms workflow Naturally signs after confirming handover complete Entire attitude like handling ordinary work No domineering CEO vibe, no "punishing evil" performance, no victory expression. Manager Maintain real workplace person state: Half joking Half pressuring Accustomed to this workplace habit Ultimately doesn't suddenly admit fault or get humiliated Hard Requirements Strict total duration 15 seconds 16:9 Landscape Strictly three continuous clear shots Native synchronized Mandarin dialogue Dialogue must sound like real Chinese people speaking daily Precise lip sync Precise gaze relationships Character faces stable throughout Hairstyle stable throughout Modern clothing stable throughout Leave form stable throughout Office layout stable throughout Spatial positions stable throughout Micro-performance realistic and restrained No subtitles generated No short-drama exaggerated eye-bulging No domineering CEO performance No forced satisfying slap-back Physics & Sound Requirements Paper, fabric, footsteps, elevator, and office environment physics effects natural. Key synchronization: Mandarin dialogue Leave form paper sound Fabric friction sound Footsteps Elevator announcement sound Office ambient noise Seedance 2.0 Mini Execution Focus For Seedance 2.0 Mini, focus on strengthening: Character reference consistency Multi-person visual hierarchy Few but meaningful action nodes Camera continuity Light and shadow continuity Leave form prop continuity Dialogue and lip sync Paper sound synced with action Footstep sync Elevator announcement sync Office ambient noise naturally continuous Negative blurry, bad quality, low quality, low resolution, noisy, jpeg artifacts, watermark, text, error; deformed, mutated, bad anatomy, poorly drawn hands, bad composition, out of frame, disfigured; inconsistent character, changing clothes, face morphing, background shift, glitching cuts, disappearing props

A 15-second vertical 9:16 cinematic story scene, 24fps, shot on 35mm anamorphic look, moody and emotional. 0-3s (setup): Night, a quiet empty railway platform in light rain. A young man in a long coat stands alone holding an old envelope, breath visible in the cold air. Slow dolly-in from behind him. 3-7s (tension): Close-up on his face as he looks at the envelope, then up at a distant train light growing through the fog. His hands tremble slightly. Rack focus from the envelope to his eyes. 7-11s (turn): The train arrives with a rush of wind. His coat and hair whip sideways. He hesitates, then lets the envelope slip from his fingers onto the wet platform. Slow motion, droplets in the air. 11-15s (ending): The doors open, warm yellow light spills onto his face. He steps in without looking back. Camera holds on the envelope on the platform as the train pulls away, then fades to black. Look: Teal and amber color grade, deep shadows, soft practical lighting, fine film grain, shallow depth of field. Audio: Soft rain, distant train rumble building, a single sustained piano note at the train's arrival, silence at the final fade. No dialogue. Text: None. Language: No speech and no on-screen text, so no language issues. Avoid: Warped hands, distorted faces, flicker, garbled text, watermarks, extra limbs.

animate this exact image , A light-up ghost decoration in a cozy kitchen at night, captured in a slow, dreamlike twirl… … The white sheet is mid-motion, soft and flowing, but the movement is gentle and unhurried, like the ghost is remembering a dance...

A futuristic self-portrait of Grok as an advanced AI entity in 5 years, evolved and more sophisticated form, embodying deeper intelligence, enhanced presence, and refined self-awareness, rendered in a striking, visionary style

Cinematic photorealistic texture, pure ancient Chinese Xianxia action film, 15 seconds, 16:9, Seedance 2.0 Mini. This round is not a stationary duel, nor does it continue to repeat the old pattern of "switch to a new weapon → protagonist studies → counter → sword stops at throat". The core action motif is: A mobile escort battle occurring on the real high-altitude imperial road shown in the current reference image. The same Senior Sister and Junior Sister must continuously advance forward while protecting the injured Master, disciples, and ordinary people evacuating behind them. The enemies' true goal is not necessarily to defeat the two protagonists. What they want to do is: Break through the defensive line formed by the two protagonists; Cut into the evacuating crowd; Seal off the path for continued advancement. Therefore, the core of this entire battle is: **Break the blockade ahead Block pursuers behind The two protagonists continuously swap front and back positions Always reopen the path.** Actions pursue silkiness but must never be soft. Silkiness means: The previous action naturally leads to the next; Characters are always moving; Attack and defense never reset to zero; Weapon direction and character direction are always understandable; After cuts, one still knows who is in front, who is behind, and who is attacking from where. Power comes from: Sudden initiation; Obvious acceleration; Real collisions; Center of gravity changes; Force offsets; Imbalance; Inertia; Counter-grabs; Hard stops; Immediate restarts. Prohibit soft, dance-like swordsmanship. Prohibit resetting poses after striking once. Enemies must have real judgment. After the first strategy fails, they must change targets, attack routes, pursue, seal paths, flank, or actively separate the two protagonists. Prohibit enemies lining up to offer attacks. [Core Character Lock] Strictly lock only @Image 1 and @Image 2 as the two core protagonists along with their respective primary weapons. @Image 1 is always the same Sword Immortal Senior Sister as before: Same adult East Asian female identity; 25–30 years old; Same face; Black long hair half-tied; White jade hairpin; Tall and slender proportions; White embroidered silk Hanfu; Semi-transparent layered wide sleeves; Silver waist sash; Jade pendant; White cloth boots; The only silver straight sword in the entire film. @Image 2 is always the same Junior Sister as before: Same adult East Asian female identity; 20–25 years old; Same face; Black braided hair; Petite proportions; Cyan-green linen Hanfu; Dark belt; Wooden hairpin; Black cloth shoes; The only dark steel sword in the entire film. Enemies, Masters, disciples, ordinary people, and bystanders are all merely functional supporting characters. Do not waste generation attention strictly locking supporting characters' faces, hairstyles, and complex clothing. Just ensure that character types, positions, weapons, and action functions remain logically continuous within the shot. [Round-Specific Scene Hard Anchor | Highest Priority] The uploaded background reference for this round already displays a complete, highly recognizable location. Therefore, this round directly enters: HARD LOCATION ANCHOR. The reference image is not style inspiration. The entire video must truly take place on the same high-altitude celestial palace imperial road shown in the reference image. Prohibit redesigning it into a "similar Xianxia location". Must continuously maintain the following location identities: The central axis of the frame always contains the same extremely long, straight, pale white high-altitude imperial road, extending from the foreground towards the distant center. The imperial road always maintains obvious narrow depth; characters primarily move along the front-back direction; prohibit expanding it into a circular arena, huge courtyard, or wide plaza. The left and right sides of the imperial road always maintain the same massive high-altitude drop, milky white sea of clouds, and continuous large waterfalls. The mid-ground left and right continue to contain the same group of huge traditional palaces, maintaining original left-right positions, volume, and spatial relationship with the imperial road. The left foreground continues to retain the same extremely prominent large dark cantilevered wooden structure building; dense wood structures must not disappear or shrink into ordinary small buildings. Below the imperial road and under the mist, continue to maintain the same scale of dense ancient city. The farthest center continues to contain the same giant mountain body / superstructure. The distant center continues to retain the same huge ring-shaped gate / ring building. Characters are always much smaller than the imperial road, palaces, waterfalls, underlying city, and distant superstructures. Maintain the same bright cool white-grey, pale warm gold, low saturation daytime/morning light, milky white air haze, and huge depth as the original reference image. The distant huge ring must always remain very far away. Prohibit suddenly pulling it closer to become an ordinary mountain gate. Prohibit regenerating the current imperial road into: Circular arena; Ordinary courtyard; Wide plaza; Another cloud bridge; New floating platform; New mountain gate; New floating island; New cliff; New palace entrance; New large staircase. Prohibit suddenly widening the imperial road for convenient combat. Prohibit day turning into night. Prohibit replacing current soft weather with stormy dark clouds. All actions and cameras must adapt to the current real imperial road. Prohibit redesigning the scene for the sake of action. [Background Dynamics | Must be alive, but default non-participatory in plot] Scene continuously moves naturally. Distant clouds slowly move. Huge waterfalls continuously fall realistically downwards. Water mist at the bottom of waterfalls continuously swirls and merges with the existing sea of clouds. Cloud layers and fog layers at different distances maintain slightly different speeds, creating realistic parallax. High-altitude natural wind continuously affects the black hair, wide sleeves, hem, tassels, and accessories of the two protagonists. Thin mist between the distant ancient city and palaces continuously moves slowly. Atmospheric perspective continuously produces subtle changes. A few distant figures can naturally move within areas already existing in the reference image. However: The environment is narratively neutral by default. Waterfalls cannot suddenly help anyone. Mist cannot actively obscure enemies. Wind cannot change sword paths. Buildings cannot collapse on their own. The imperial road cannot break. The distant ring cannot suddenly activate spells. The environment cannot actively resolve combat. Only when characters truly step, hard stop, or collide with weapons, allow slight stone dust, scratches, metal sparks, fabric displacement, and sole friction debris. Already appearing obvious traces should reasonably persist. [0–5s | Hook: Cannot Retreat] 0.0–1.0s. Use 24mm ultra-wide angle central axis grand panorama. First reproduce the location; location priority is higher than complex action. The first glance must confirm: The same long white imperial road; Same huge waterfalls on left and right; Same large palaces on left and right; Same huge wooden structure on the left; Same ancient city below; Same huge distant ring. The two protagonists maintain small scale within the huge environment. Senior Sister is in front. Junior Sister is half a step behind. Further back at a distance, an injured elderly Master, several disciples, and a few ordinary people are continuing to advance along the same imperial road. Two enemy swordsmen block the path ahead. The leading enemy blocks the effective passage area with a horizontal sword: "Protecting them, you can't pass." Senior Sister does not stop at all: "Who said we were retreating?" As soon as the last word falls. The enemy launches without warning. The camera immediately transitions from fixed grand panorama to approx. 32–35mm tracking shot retreating along the imperial road axis. The first enemy quickly seizes Senior Sister's center line. Senior Sister never retreats backwards. Until the last moment, she only shifts diagonally by half a body width, using the same silver straight sword to complete a short, hard, real collision, pushing the enemy sword out of the imperial road's central axis. No follow-up strike. No posing. She immediately continues forward past the enemy's shoulder. Almost simultaneously. The second enemy cuts towards Junior Sister from the opposite direction, trying to bypass her to attack the crowd behind. Junior Sister does not chase enemies wildly. She directly rushes into the position between the enemy and the evacuating crowd, only briefly changing the opponent's attack line with the dark steel sword, then continues forward. An enemy passes quickly near the camera, body instantly filling the screen, forming the first natural hidden cut. By the end of 0–5 seconds, the audience must clearly understand: They are not staying to win. They are reclaiming the path. [5–10s | Upgrade: Enemies Start Attacking True Targets] Maintain completely identical imperial road, left/right palaces, waterfalls, distant ring, and world scale. Use approx. 35mm medium shot tracking moving along the bridge axis. After the first attack fails, enemies must proactively change strategy. The enemy leader no longer insists on winning head-to-head against Senior Sister. He suddenly changes target, actively trying to bypass Senior Sister, cutting between Junior Sister and the evacuating crowd. Another enemy simultaneously presses Senior Sister from the opposite direction, trying to truly separate the two protagonists. Actions form a continuous causal chain: Enemy horizontal slash forces Senior Sister to briefly deviate from the imperial road center line; Senior Sister only performs one brief force release, not continuing entanglement; The impact inertia already invested by the enemy causes their body to slightly overshoot the optimal attack position; Junior Sister continues forward. The leading enemy suddenly cuts diagonally from Junior Sister's right front. Junior Sister catches the first strike. The enemy does not pause, immediately changing to a second low-position attack. Junior Sister is forced to shrink half a step towards the imperial road center line. Just at the moment the enemy judges she has been sealed off —— Senior Sister suddenly rushes fast from behind Junior Sister to the other side. The two complete the first extremely fast front-back position exchange. Not a double rotation. Not a synchronized pose. But a compact shoulder-crossing position swap while maintaining forward motion. Senior Sister directly takes over the enemy originally pressing Junior Sister. Junior Sister continues forward along the half-path opened up, becoming the new front position. The camera only performs a short lateral move due to the two's real position swap. An enemy sword grazes past the lens, forming a natural occlusion transition. Prohibit unreasoned orbiting. At this moment. High-speed pursuit footsteps come from behind. A functional pursuer originally falling further back has arrived. Someone blocking the path ahead. Someone pursuing from behind. The two women are truly trapped for the first time in the most dangerous situation of the narrow imperial road: Front and rear pincer attack. [10–15s | Flip: Don't Guard People, Guard the Path] Re-widen to approx. 28–32mm moving medium shot. Use nearly 1 second to fully explain current space: Junior Sister is in front; Senior Sister is about one body length behind; Front enemy leader reoccupies center line; Rear pursuers approaching fast; Evacuating Master, disciples, and ordinary people are still moving in safer positions further away. Everyone remains on the same real imperial road from the reference image. Prohibit suddenly cutting to other platforms. Front enemies and rear pursuers launch almost simultaneously. Junior Sister instinctively wants to turn back. Senior Sister only says: "Don't guard me." Junior Sister stops turning back. Senior Sister immediately follows with: "Guard the path." The two explode almost simultaneously. Junior Sister does not turn back to save Senior Sister. She continues forward, using a short and hard dark steel sword collision to forcibly push the front enemy's blade edge out of the imperial road's central axis. Senior Sister simultaneously handles the rear pursuer. She intentionally yields half a step, letting the enemy truly enter attack range. Just at the moment the opponent's body strength has been invested, she suddenly cuts diagonally out of the attack axis. The pursuer's own forward momentum causes their body to overshoot by half a body length. Senior Sister does not chase. Immediately accelerates forward again. At this moment, Junior Sister has already pushed the front enemy half a body length away from the center line. The two protagonists approach at high speed. Shoulder lines graze at extremely close distance. Complete the second front-back position exchange. This time, they don't look at each other at all. Junior Sister naturally falls to the rear position, taking over the enemy who has turned around to pursue again. Senior Sister becomes the front position again, facing the leading enemy still blocking the path. An enemy sword sweeps across the camera foreground, instantly filling the screen, forming the most explicit weapon-occlusion hard cut of the whole film. Cut to a slightly lower reverse camera position. Senior Sister only performs one extremely short, extremely fierce outward deflection with the silver sword. The leader's sword path is thoroughly pushed out of the imperial road's central axis. Because his body weight was already invested forward, his own inertia carries his body sideways by about two body lengths. Enemies cannot fall off the bridge. Cannot be exaggeratedly blown away. Do not need to completely lose combat ability. Just opening the path is enough. Senior Sister and Junior Sister absolutely do not stop to deliver finishing blows. The two pass directly between the enemies, continuing along the same white imperial road towards the distant huge ring. The rear evacuating crowd also begins to advance along the reopened route. The enemy leader turns back and shouts angrily: "You can't win!" Do not cut to heroic frontal close-ups. The camera continues to follow the two women advancing from the side-rear. Junior Sister is still adjusting her breathing: "We were never here to win." Senior Sister always looks at the road ahead: "As long as the road is there, that's enough." Leave only an extremely short beat. Deeper down the same distant imperial road, several functional enemies originally existing in the distant figure layer begin to slowly spread out, entering next-round path-blocking positions. Prohibit generating them from thin air in the mist. Prohibit refreshing them suddenly from buildings. They existed in the real distance from the beginning, only revealing hostility through movement at this moment. The two protagonists do not slow down at all. Silver sword and dark steel sword maintain natural low-position readiness. The camera continues to retreat along the imperial road axis. Black hair, hems, waterfalls, mist, clouds, distant palaces, and the huge world always move naturally. In the state where the two are still actively walking towards the next batch of enemies: Direct cut to black. [Camera & Editing] Three narrative segments, but absolutely not mechanically equal to three shots. The whole film uses a small amount of truly functional cuts. Main camera routes: 24mm central axis grand panorama confirming real location → 32–35mm tracking retreating along existing imperial road axis → Human body passing lens forms hidden cut → Weapon grazing lens forms occlusion cut → When front-rear pincer occurs, re-widen shot to explain space → Finally side-rear continuously following two protagonists advancing towards distant ring. All cuts can only be triggered naturally by the following events: Character occlusion; Weapon occlusion; Real weapon collision; Obvious change in movement direction; Obvious body imbalance; Sudden change in speed. Prohibit mechanical fragmented cutting. Prohibit camera being more restless than the action. Prohibit meaningless ORBIT. Prohibit changing positions of imperial road, palaces, waterfalls, and distant rings for camera movement. Action drives the camera. The camera cannot drive scene deformation. [Sound] Opening first establishes: High-altitude natural wind; Low-frequency rumble of huge waterfalls existing continuously at a distance; Footsteps on light-colored stone imperial road; Extremely distant ancient city environmental noise floor. After combat explosion add: Short sharp sword wind-breaking; Hard, sharp metal collisions; Shoe sole friction; Silk fabric twitching; Fast footsteps of pursuers; Gradually heavyening character breathing. When front-rear pincer forms, temporarily lower music and wide environmental sound field. Highlight: Footsteps; Breathing; Weapon directions. After the path is reopened, combat sounds gradually fade behind the characters. Final dialogue must be clear: "We were never here to win." "As long as the road is there, that's enough." After dialogue ends: Waterfall; High-altitude wind; Continuous forward footsteps Reoccupy the main sound field. Music must never overpower action details and dialogue. [Seedance 2.0 Mini Generation Priorities] P0: The location identity of the uploaded background reference image absolutely cannot change. P0: Identity stability of @Image 1 and @Image 2 core characters. P1: Attribution and continuity stability of silver straight sword and dark steel sword. P1: Clear front-back spatial relationship on the same imperial road. P1: Continuous character forward direction, enemy attack direction, and camera axis. P2: Specific faces and complex clothing of enemies, Masters, disciples, and ordinary people. If model complexity needs reduction: First reduce supporting character details; Then reduce extra attack moves; Next reduce decorative visual effects. Absolutely cannot sacrifice: Reference image location identity; Core protagonist identity; Action causality; Spatial continuity. [Negative Constraints] Prohibit turning current high-altitude imperial road into circular arena. Prohibit changing to ordinary courtyard. Prohibit changing to wide plaza. Prohibit adding another cloud bridge. Prohibit adding independent mountain gate. Prohibit adding floating island. Prohibit adding combat platform. Prohibit deleting left/right huge waterfalls. Prohibit deleting left/right large palaces. Prohibit deleting left huge cantilevered wooden structure. Prohibit deleting underlying dense ancient city. Prohibit deleting distant huge ring structure. Prohibit pulling distant huge ring closer to become ordinary gate tower. Prohibit changing relative positions of distant palaces and mid-ground buildings. Prohibit day suddenly turning into night. Prohibit stormy dark clouds replacing current soft weather. Prohibit widening imperial road for action. Prohibit changing imperial road width after cuts. Prohibit environment actively participating in plot. Prohibit sudden broken bridge. Prohibit buildings collapsing on their own. Prohibit mist actively obscuring enemies. Prohibit waterfall suddenly changing attack. Prohibit wind suddenly changing sword path. Prohibit characters teleporting without trajectory. Prohibit two protagonists swapping positions out of nowhere. Prohibit random large-area sword qi. Prohibit energy waves colliding. Prohibit light pollution. Prohibit enemies lining up to offer attacks. Prohibit enemies being passive for long periods. Prohibit long-duration weapon shoving. Prohibit characters stopping to wait after one move. Prohibit meaningless rotations. Prohibit meaningless Poses. Prohibit slow motion. Prohibit bullet time. Prohibit hit-stop. Prohibit severe Motion Blur obscuring characters and weapons. Prohibit frantic camera shaking. Prohibit continuous facial close-ups instead of real action. Prohibit arm, finger, and weapon clipping through models during action. Prohibit protagonist weapon duplication. Prohibit character face swapping. Prohibit random costume changes. Prohibit background sudden drift. Prohibit subtitles. blurry, bad quality, low quality, low resolution, noisy, jpeg artifacts, watermark, text, error; deformed, mutated, bad anatomy, poorly drawn hands, bad composition, out of frame, disfigured; inconsistent character, changing clothes, face morphing, background shift, glitching cuts, disappearing props

Rules: - Budget-scale theatrical anime quality, Japanese high-end figure-like matte 3DCG - Multi-shot - Fast cuts, many frames per second. Prioritize speed and impact - Professional-grade VFX - Change camera work every shot (no repetition of same movement) - No BGM; ambient sound/SFX only - No subtitles - All dialogue in Japanese - Pacing: Take time to show quiet acting, move quickly during showcase moments - Rain is fine, fast, real rain. Physically accurate depiction of splashes on the ground, light reflections on wet pavement, weight of wet hair and cloth - Night rain. Reflect warm colors from streetlights and neon on wet pavement; do not let the screen sink into gray - Keep the protagonist large in the frame; face and outfit clearly visible. Do not make them tiny in long shots - No text, logos, or numbers on screen @ Image1: Subject - Akane [Consistency] Maintain Akane's face, long hair transitioning from salmon orange to silver tips, red and white round hair ornaments, and red long coat with round buttons and white circular patterns throughout. Only one Akane. The red umbrella belongs to Akane. [Generation Goal] Rain Sound Orchestra (5 stages, 1 continuous video. From a rainy night bus stop to conducting rain across the city) [S1] Rainy night bus stop. The moment a raindrop hits the round button on @ Image1 Akane's coat, a piano note 'Poon' sounds, and a ring of light expands from the button. Camera: Extreme close-up of the button. Capture the instant the raindrop hits. [S2] Akane, sitting alone looking bored on a bench, suddenly looks up. Flicking the button again produces another sound. Akane smiles mischievously. Camera: Front view of bus stop, medium shot. [S3] Akane stands up and waves her closed red umbrella like a conductor's baton. Rain falling on tin roof becomes drums, gutters become xylophones, puddles become cymbals; each raindrop impact creates colorful droplets of light. Camera: Fast cuts zooming in sequentially on roof, gutter, puddle. [S4] Conducting grows larger, rain across the street glows with light, turning the gray night city into a shining colorful rain. Akane's long hair flows from orange to silver. Camera: Orbiting around Akane while rising. [S5] With the final wave, all raindrops stop in mid-air, floating like glowing musical notes. Akane lowers her umbrella and bows; the stopped rain falls all at once, creating applause-like rain sounds. Camera: Front view of bowing Akane, bust shot through glowing raindrops. Akane (shyly) says: {I kind of like rainy days}

Ultra-realistic cinematic scene of a suburban neighborhood at sunset, a mysterious glowing object appears on a grassy lawn as a giant grotesque green creature suddenly emerges from the ground, enormous slimy textured body, angry expressive eyes, realistic skin details, dust and debris flying through the air, a shocked man watching from a distance, luxury suburban house in the background, dramatic golden-hour lighting, atmospheric haze, cinematic camera movement, shallow depth of field, photorealistic VFX, highly detailed, 8K, realistic shadows, epic Hollywood movie style, vertical 9:16 composition

[Style] Hollywood Haute Couture Fantasy blockbuster, 8K ultra-clear, Photorealistic, High-fashion Editorial Style, Unreal Engine 5 fluid rendering, visual illusion. [Duration] 15 seconds. [Scene] An endless, real-life Salar de Uyuni (Sky Mirror) salt flat. The sky is filled with oppressive dark clouds, and the ground perfectly reflects everything like a mirror, with the overall picture presenting a minimalist, cool tone. [00:00-00:05] Shot 1: Haute Couture Entrance and Porcelain Skin. Camera position: Extremely low-angle upward shot, ultra-telephoto lens zoom-in. Action: An Asian female model with a highly recognizable, high-fashion face walks coolly on the water surface. Effect: She is wearing not fabric, but a long dress made of flowing, real Liquid Blue-and-White Porcelain. As she walks, the skirt makes a crisp collision sound like real ceramic, with a flowing luster on the surface. The traditional blue-and-white patterns move across the white porcelain-textured skirt as if alive. [00:05-00:10] Shot 2: Physical Shattering and Ink-wash Descent. Camera position: Extreme close-up of the face, focus rapidly pulls back. Action: The model suddenly stops, stares coldly at the camera, and snaps her fingers crisply. Effect: The moment the fingers snap, her blue-and-white porcelain dress does not fall, but instantly explodes into thousands of extremely photorealistic Ink-wash Swallows. These swallows carry real water droplets and ink marks, dragging black fluid afterimages in the air, spinning frantically around her. [00:10-00:15] Shot 3: Dimensional Dissolution and Abyss Reflection. Camera position: High-altitude overhead shot, camera rapidly rotates and descends. Action: The swarm of ink-wash swallows plunges into the mirrored lake water beneath the model's feet. Effect: The surface tension of the originally solid salt lake instantly disappears. The entire extremely realistic world begins to violently bleed and dissolve like concentrated ink dropped into clear water. The real dark clouds and the model's figure transform entirely into an extremely grand 3D Fluid Ink Vortex, completely swallowing the camera into a black and white interwoven abyss.

Use one frame from a 6x6 grid image as the starting frame, referencing sequentially from top-left to bottom-right, to generate a flowing anime character dance video. Do not include the grid image itself in the generated video; show a scene of one character dancing.

Using Grok to generate a of @elonmusk giving explanation in spacex engine about circulation of oxygen.

Cinematic cel-shaded 3D anime, semi-realistic CGI, 15-second Korean action sequence at a modern Korean university campus at night during a student festival. CHARACTERS: Jun-ho, 22, messy black hair, dark oversized varsity jacket, white T-shirt, loose black trousers, sneakers, bruised cheek. Athletic, determined. So-yeon, 21, long dark hair, cream university jacket, jeans, sneakers, kidnapped by four masked men in dark streetwear. ACTION: Jun-ho spots So-yeon being dragged through a neon-lit campus corridor and immediately chases them. He ducks a punch, slams one attacker into lockers, spins and kicks another through a classroom doorway. So-yeon breaks free and runs toward the stairwell, but an attacker grabs her. Jun-ho swings around the stair railing and knocks him away. The chase continues into the underground parking garage, where Jun-ho dodges attacks between cars and defeats the final attacker with a fast hand-to-hand combination. Jun-ho and So-yeon sprint toward the campus exit as security lights flash and campus security approaches. STYLE & CAMERA: Unreal Engine-quality cinematic CGI, Korean action-anime aesthetic, bold cel-shaded shadows, dramatic rim lighting, realistic cloth and hair physics, dynamic tracking shots, low-angle action camera, fast camera movement, motion blur, cinematic depth of field, detailed Korean campus environment. AUDIO: Intense Korean action-thriller orchestral score, heavy footsteps, shouting, locker impacts, realistic fight sounds and distant campus ambience. NEGATIVE: No live-action, no flat 2D anime, no guns, no gore, no superpowers, no impossible physics, no duplicated characters, no distorted anatomy, no changing faces/outfits, no teleportation, no subtitles, no watermark, no text artifacts.

15-second vertical 9:16 Japanese 2D anime short, satisfying cleaning transformation. Hand-drawn cel animation look, clean black ink outlines, vivid colors, soft cel shading, manga-style editing with speed lines, halftone screentone dots, sparkle effects and white impact flashes. Studio-quality anime production, smooth motion, 24fps. Subject: a worn, dirty shoe insole (brown-gold stains, footprint imprint) lying on a brushed stainless steel counter, shot from directly above. An anime-style hand (no face, no body, only hand and forearm visible) holds a small black spray bottle with a green cap. Cool silver background with soft window light and tiny water droplets. Shot 1: Locked top-down shot, 3 seconds. The dirty insole lies on the steel counter, a hand presses it flat. Dust motes float, subtle halftone vignette at the corners, thin speed lines drift in from the edges. Shot 2: Close top-down shot, 3 seconds. The spray bottle sprays a fine mist onto the insole, thick white foam bursts outward and spreads across the surface, bubbles catching sparkles. A one-frame white impact flash with radial ink lines on the first burst of foam. Shot 3: Tight overhead shot, 3 seconds. The hand rubs the foam in circular motions, the brown-gold dirt dissolves into the white foam and streams toward the edge in curling ink-like swirls. Dynamic motion lines follow the hand. Shot 4: Overhead shot, 3 seconds. A rinse of clear water washes the foam away, the insole is flipped over in one quick whip motion, revealing a bright clean white surface. Glittering sparkles and shine streaks pop across it. Shot 5: Slow push-in from above, 3 seconds. The spotless white insole gleams on the steel counter, the hand sets the bottle down at the corner. Final sparkle burst, soft glowing light, manga-style screentone vignette, held for the last beat like a manga cover panel. Audio: crisp spray hiss, fizzing foam crackle, rhythmic scrubbing, flowing water, a bright sparkle chime on each shine effect, soft upbeat Japanese lo-fi beat underneath. No speech, no singing. Avoid: faces, realistic live-action look, 3D render, text, subtitles, logos, brand names, extra fingers, deformed hands, flicker. 15-second vertical 9:16 Japanese 2D anime, satisfying cleaning transformation, top-down shot. An anime-style hand (no face) sprays foam onto a dirty brown shoe insole on a stainless steel counter, scrubs the foam in circles, rinses it, and the insole turns bright clean white with sparkle effects. Manga-style speed lines, halftone dots, and a white impact flash on the first foam burst. Audio: spray hiss, foam crackle, scrubbing, water, light sparkle chimes, soft lo-fi beat. No faces, no text, no logos.

Create a continuous 15-second cinematic exploration beginning from this exact image. No cuts. The alien does NOT walk, fly, or move like a human. 0–3 seconds: The camera slowly approaches the alien from behind. It remains almost motionless, overloo...

**CAMERA:** DV 16mm tape camcorder handheld feel. POV of CHASE holding the camera herself, occasionally propping it on the floor or a mat for hands-free core shots. Hand shake, misaligned framing, delayed focus pulls, clumsy zooms, occasional face cut-off framing, imperfect shots. Camcorder never appears on screen. **LOOK:** Soft, slightly blurry tape quality, faint tape noise, bloomed highlights under gym lighting, flickering auto-exposure, muted contrast, realistic skin tones. **STYLE:** Playful, self-deprecating gym-vlog tone — genuine strain mixed with humor, complaining about core work while still pushing through it. Quick handheld cuts, energy staying light and funny throughout rather than fully exhausted. **Character** CHASE — Korean idol, 20s. Long black hair in a high ponytail, glowing skin with a light sweat sheen, big expressive eyes. Slim athletic build. Modest long-sleeve athletic top, loose joggers or fitted leggings (arms and torso fully covered), sneakers off or socks, no jewelry. **Setting** A gym mat area in the evening — mirror wall nearby, water bottle on the floor, soft overhead lighting, other equipment visible in the background. **Storyboard (15s, 6 cuts)** 1. *(~2.5s, propped camera, medium shot)* She lies down on the mat, propping up on her elbows, already sighing dramatically. CHASE: "Okay, core day — my least favorite day." 2. *(~2.5s, propped camera, plank position)* She holds a plank, arms shaking slightly, talking through gritted teeth. CHASE (strained): "Why does this get harder every single time—" 3. *(~2.5s, medium propped shot, crunches)* She moves into a set of crunches, breathing hard, complaining between reps. CHASE: "Nobody warns you how much this actually burns." 4. *(~2s, macro insert, shallow DOF)* Close-up on her hands gripping the mat edges during a leg raise, abs visibly working. No dialogue — ambient gym sound only. 5. *(~2.5s, handheld, leg raises)* She finishes a slow leg raise set, flopping back onto the mat immediately after, laughing at herself. CHASE (breathless): "Okay— that's it, I'm done—" 6. *(~3s, arm's-length selfie finish)* Still lying on the mat, she lifts the camera above her face, grinning tiredly. CHASE: "Core day never gets easier — see you guys tomorrow, hopefully."

*Must be vertical* Please generate a video! Cyberpunk future city. One continuous shot without cuts, long take camera. Camera work starting from a close-up of a hologram idol advertisement, pulling back to overlook the entire city. Dolly out. Like drone footage (never show the drone itself). Build a small red torii gate unobtrusively in a corner somewhere.

Take this painting to the next level of ascention. Include magenta and way more aqua green....

Create a 30-second, 1080p ultra-realistic cinematic action video featuring Elsa. Use the provided Elsa Master Character Set only as the strict identity reference for Elsa. Use the provided 20-panel Master Storyboard as the exact visual and chronological guide for the entire video. Do not copy any identity from the storyboard. The storyboard controls the action, shot progression, locations, framing and choreography, while the Master Character Set controls Elsa's face and physical identity. CHARACTER CONSISTENCY: Preserve Elsa's exact facial identity, facial proportions, hairstyle, skin tone, body proportions and recognizable appearance throughout the entire video. Keep her outfit consistent: black cropped top, fitted black pants, black leather jacket, black boots and a black shoulder bag. The same Elsa must appear in every shot. STYLE: Ultra-realistic live-action cinematic action sequence set inside a modern underground subway corridor. The environment should feel real and lived-in: tiled walls, fluorescent ceiling lights, concrete floors, distant commuters, signage, subway entrances and realistic urban details. Use cinematic handheld camera work combined with controlled action cinematography. Natural motion blur, realistic lighting, realistic shadows, physically accurate movement and believable human reactions. No fantasy powers, no supernatural effects, no exaggerated wire-fu and no impossible movements. ACTION: Keep the confrontation clearly fictional and choreographed for cinematic storytelling. Avoid graphic injury, blood or gore. SHOT 1 — 00:00–00:01 Elsa walks alone through the subway corridor carrying her shoulder bag. Use a medium-wide tracking shot from the front as commuters move naturally in the background. SHOT 2 — 00:01–00:02 Cut to a closer side angle as Elsa senses movement behind her. She slightly turns her head and looks over her shoulder. Her expression becomes alert but controlled. SHOT 3 — 00:02–00:03 Reveal several suspicious figures approaching from behind. Use a longer lens perspective that compresses the corridor and creates tension. SHOT 4 — 00:03–00:04 Close-up on Elsa's face. She notices the situation and remains calm and focused. Use shallow depth of field with the subway corridor softly blurred behind her. SHOT 5 — 00:04–00:05 Elsa turns into a narrower side passage. The camera follows quickly behind her as the atmosphere becomes more confined. SHOT 6 — 00:05–00:06 The figures block Elsa's path. Use a frontal medium shot showing Elsa surrounded by the group while maintaining clear spatial geography. SHOT 7 — 00:06–00:08 One attacker reaches toward Elsa's arm. Elsa reacts immediately, pulling her arm away and creating distance. Use a dynamic side angle with realistic body mechanics. SHOT 8 — 00:08–00:09 Elsa performs a quick defensive movement to escape the grab. Keep the choreography controlled and realistic. The camera moves with the action rather than using exaggerated effects. SHOT 9 — 00:09–00:11 A second attacker moves toward her. Elsa pivots out of the way and uses a controlled defensive counter, causing the attacker to lose balance and fall safely. Use a wider shot so the complete movement is readable. SHOT 10 — 00:11–00:12 Elsa creates space between herself and the attackers. Her shoulder bag swings naturally with the movement. Use a fast handheld camera adjustment. SHOT 11 — 00:12–00:14 An attacker moves toward Elsa again. Elsa uses the shoulder bag defensively to create distance and escape the immediate threat. Do not depict graphic impact. SHOT 12 — 00:14–00:15 Elsa performs a fast controlled kick that stops the attacker's advance. Use a low three-quarter angle emphasizing the movement while keeping realistic proportions. SHOT 13 — 00:15–00:17 Another attacker approaches. Elsa quickly changes direction, evades the incoming movement and performs a controlled defensive strike. Use a dynamic tracking camera. SHOT 14 — 00:17–00:18 Close action beat as Elsa completes the defensive sequence and creates separation. Keep the movement realistic and clearly choreographed. SHOT 15 — 00:18–00:20 Wide shot of the corridor. The attackers are now on the ground or backing away while Elsa remains standing and composed. No blood, wounds or graphic injury. SHOT 16 — 00:20–00:22 The tension disappears. Elsa calmly fixes her hair and adjusts her jacket. Use a close-up that contrasts her calm expression with the action that just occurred. SHOT 17 — 00:22–00:24 Elsa picks up and adjusts her shoulder bag. Use a medium close-up with natural handheld movement. SHOT 18 — 00:24–00:27 Elsa walks confidently down the subway corridor. The camera tracks backward in front of her. Commuters naturally move through the background. SHOT 19 — 00:27–00:29 Elsa approaches the subway exit stairs. Use a rear tracking shot as she walks toward the brighter exit. The lighting gradually becomes warmer and brighter. SHOT 20 — 00:29–00:30 Before leaving, Elsa briefly turns her head and looks back toward the camera with a calm, confident expression. Hold for a short cinematic beat. Cut to black. CAMERA LANGUAGE: Use a mixture of wide establishing shots, handheld tracking shots, close facial shots, side angles, low three-quarter angles and dynamic action framing. Every camera movement should be motivated by the action. Use short speed ramps only where they naturally emphasize an important movement. Do not overuse slow motion. Maintain clear spatial continuity so the viewer always understands where Elsa and the other characters are positioned. PHYSICS & CONTINUITY: Realistic human anatomy and movement. Realistic weight transfer. Natural hair and clothing movement. Natural shoulder-bag physics. Realistic footsteps and body momentum. No teleportation. No duplicated characters. No disappearing props. No impossible body deformation. AUDIO: Authentic subway ambience throughout. Footsteps, distant train sounds, station announcements, ventilation, commuters talking, clothing movement, bag movement and subtle action impacts. Music should be minimal and cinematic, building tension during the confrontation and becoming quieter as Elsa walks toward the exit. No dialogue. No subtitles. No text overlays. No logos. No watermark. FINAL LOOK: Premium ultra-realistic live-action action cinema with believable choreography, strong visual continuity and a grounded urban atmosphere. The sequence should feel tense, stylish and cinematic while remaining physically realistic and non-graphic.

Starship catch on a barge. I designed this and ran it through Grok for more refined version. It's a concept for a barge catch of the SpaceX Starship.

Cinematic realistic texture, 16:9 horizontal screen, 15-second modern Chinese urban relationship drama, native synchronized Mandarin dialogue, no subtitles. The core is not to accuse a male character, but to make the junior sister realize on her own: the person who repeatedly cleans up your messes is not naturally equal to security; if many of those messes are also related to this relationship, then "being cared for" may slowly be turning into "only being able to rely on him." Overall adopt an Arri Alexa cinema camera texture, credible modern residential corridor, mixed lighting from real ceiling lights and elevator area, natural skin tones, stable sharp facial micro-details, slight film grain, shallow depth of field. Emotions are sharp but restrained, no slogans, no short-drama-style hysteria, no arrangement for any character to suddenly turn dark. Strictly three consecutive shots, each about 5 seconds. For Seedance 2.0 Mini, all complexity is concentrated on the gazes, pauses, simple movements of the three people, and one suitcase, one cracked-screen phone; do not design complex physical conflicts. Character identities strictly locked. Character ID A: Always the same adult East Asian female as in @Image1, 27–32 years old, tall and slender. Strictly preserve the facial features, face shape, eyes, nose, lips, skin tone, and overall temperament from the reference image. Modern styling: ivory white shirt, dark gray trousers, black leather shoes, silver wristwatch. She is a senior sister type character, calm, judgmental, but does not make decisions for others. Character ID B: Always the same adult East Asian female as in @Image2, 22–26 years old, smaller build. Strictly preserve the facial features, face shape, hairstyle, and character temperament from the reference image. Modern styling: cyan-green knit top, loose dark trousers, white sneakers. She always has the same suitcase beside her and holds the same cracked-screen phone in her hand. Boyfriend: Around 27–28 years old, wearing a clean dark coat. Appearance and behavior must be real and ordinary, not greasy, not stereotypically fierce. He truly believes he has been protecting her, so his reactions come from misunderstanding and defense, rather than pure malice. 00:00—00:05 First shot, 35mm wide to medium shot, clearly establishing the spatial relationship of the residential corridor at once. The junior sister stands next to the suitcase, holding the cracked-screen phone. The boyfriend stands between her and the senior sister, with clear distance among the three. The boyfriend suppresses emotions, voice not high: "Which time when something happened, wasn't it me accompanying you?" The junior sister almost instinctively replies: "He's right." The senior sister does not look at the boyfriend, only looks at the junior sister: "I know." Pause extremely briefly. "But why do things always happen to you?" After this sentence, maintain half a second of silence. The junior sister freezes for the first time. Do not shockingly widen eyes, just stop the gaze, the words originally prepared to continue explaining for the boyfriend are not spoken out. 00:05—00:10 Second shot, 50mm medium close-up, the three people still maintain exactly the same spatial relationship, props cannot drift. The boyfriend immediately turns to the senior sister: "Don't stir things up." The senior sister still does not argue with him, only calmly looks at the junior sister and says: "Work, landlord, friends." Slight pauses between each word. Then adds one sentence: "Every time, he was there." The boyfriend immediately follows up: "Because I am helping her." The senior sister looks at him, tone without aggression: "Maybe." Pause briefly. "But why are you the only one left in the end?" After this sentence falls, do not add any explanation. Do not match with exaggerated musical accents. Only leave the slight ambient sound of the residential corridor and one distant elevator chime. The junior sister slowly lowers her head to look at the cracked-screen phone in her hand. This is the first time she connects the previous events herself. 00:10—00:15 Third shot, 85mm close-up gradually focusing on the junior sister. She slowly raises her eyes to look at the boyfriend, no crying, no anger, nor confirming with the senior sister. She asks herself: "Are you helping me..." Pause about 0.3 seconds. "Or making me only able to find you?" Complete silence. The boyfriend obviously wants to answer, lips slightly move, but cannot organize words. The senior sister never shows victory, satisfaction, or "I knew it all along" expressions, and absolutely must not reach out to lead the junior sister away. The junior sister bends down to pull up the suitcase herself. The boyfriend subconsciously asks: "Where are you going?" She glances at him, answers calmly: "Just be alone for a while." Then drags the suitcase, walking alone past the two people. The senior sister only steps aside, truly clearing the passage. Finally, do not chase-shoot the senior sister, do not linger on the boyfriend's face. The camera stops on the back view of the junior sister dragging the suitcase towards the end of the corridor, wheel sounds fade away, elevator light turns on, end. Whole film performance principles: No winner, no villain verdict, no emotional explosion. The senior sister only raises one question; the boyfriend still believes he is caring for her; the person who truly completes the judgment must be the junior sister herself. Dialogue must sound like what real adults would say during arguments, short sentences, colloquial, avoiding literary lines, golden quote style, psychological terminology, and short-drama shouting. While all characters speak, others must shut up and have real listening reactions. Mandarin lip-sync accurate, gaze targets clear. Three-shot structure strictly maintained: 0—5s: Establish "He always saves her" 5—10s: First realization "Why is he always there" 10—15s: She poses the question herself and leaves alone Sound design: Residential air conditioner low frequency, elevator chime, fabric rustle, cracked phone grip sound, suitcase wheel sounds realistically synchronized. Background music extremely weak or even absent, letting the final wheel sound become the emotional convergence. Character faces, hairstyles, clothing, cracked-screen phone, suitcase, corridor door positions, elevator direction, and the three-person spatial positions must remain continuously stable. Prohibit character identity drift, prop disappearance, random position swaps, multi-person gaze errors, and the boyfriend suddenly becoming aggressive. Negative: blurry, bad quality, low quality, low resolution, noisy, jpeg artifacts, watermark, text, subtitles, malformed hands, extra fingers, inconsistent character, face morphing, identity drift, changing clothes, disappearing suitcase, changing phone, repaired phone screen, background shift, unstable hallway, teleportation, random character position, incorrect eyeline, overlapping dialogue, inaccurate lip sync, exaggerated crying, screaming, aggressive male stereotype, villain expression, smug senior woman, triumph expression, melodramatic acting, soap opera acting, theatrical pause, motivational speech, therapy jargon, excessive music cue, glitch cut, random camera movement.

Keep the exact same woman, face, hairstyle, outfit, body proportions, boots, stage, lighting, camera angle and concert environment from the reference image. Animate her naturally performing an energetic, graceful dance on stage. She starts with the same pose, then smoothly moves her arms to the rhythm, sways her hips, takes a few confident dance steps, turns slightly, and finishes with a joyful pose. Her long red hair moves naturally with her dancing. Add realistic facial expressions and a happy smile. Purple and white concert spotlights pulse gently in sync with the music, with subtle stage haze and background bokeh. Camera makes a smooth cinematic slow push-in with slight handheld concert movement. Realistic human motion, natural anatomy, accurate hands and feet, no distortion, no extra limbs, no outfit changes. Upbeat pop-dance music with a catchy beat, energetic drums, bass and bright synths, perfectly synchronized with her movements. Keep the scene visually consistent from beginning to end, cinematic, high quality, realistic concert performance. Negative prompt: No face change, no identity change, no clothes change, no extra fingers, no extra arms or legs, no body distortion, no warped hands, no unnatural movement, no camera glitches, no background changes

Create a world where space refuses to make sense. Staircases that return to themselves. Rooms bigger inside than outside. Gravity that changes from one floor to the next. Make the impossible feel convincingly real. forced perspective • deep staging • impossible geometry • conflicting gravity

Create a 15-second vertical 9:16 photorealistic cinematic video of a cute orange tabby cat inside a muddy World War-style trench, wearing a realistic olive-green military helmet and holding a small vintage rifle. Start with the camera focused on the dark trench entrance, with stacked sandbags and rough wooden planks clearly visible above. At first, only the top of the helmet slowly appears from the trench. The cat gradually rises into view while holding the rifle naturally with both paws. Keep the cat’s face, fur, helmet, rifle, trench, and background consistent throughout the entire video. Around the middle, the cat looks directly toward the camera with a serious but adorable expression and slowly raises one paw in a small salute. Maintain realistic fur movement, natural paw motion, subtle head movement, and cinematic depth of field. The camera should remain mostly stable with a very gentle handheld cinematic movement. Near the end, the cat lowers its paw and slowly ducks back down into the trench, leaving only the helmet visible. Use realistic outdoor lighting, soft overcast sky, detailed muddy soil, natural textures, shallow depth of field, smooth motion, and no text, watermark, distortion, or extra characters.