Stadtbild / Urban

Create a 30-second high-quality 3D cinematic animated short about a young newspaper delivery girl trying to complete her morning route, but the entire city seems determined to make her late. She wears a bright yellow delivery jacket, dark pants, sneakers, and a delivery backpack filled with newspapers. Expressive face, charming animated-film aesthetic, detailed environments, realistic 3D animation. Sunrise over a busy city. The girl wakes up, checks her watch, gasps, grabs a stack of newspapers and rushes outside with her bicycle. She jumps onto the bicycle and races through the streets, delivering newspapers to houses while checking the growing pile in her backpack. A huge traffic jam blocks the road. She looks frustrated, quickly turns around, and squeezes through a narrow side street. A city bus suddenly blocks her path. She brakes hard, jumps off the bicycle, runs around the bus, then hops back on and speeds away. She reaches a crowded intersection. Pedestrians cross everywhere, a dog runs across her path, and a delivery truck blocks another route. She dodges everything with quick, comedic movements. She spots a shortcut through a lively market. She races between stalls, narrowly avoiding boxes and carts while newspapers fly slightly from her backpack. She arrives at the final apartment building, jumps off the bicycle, grabs the last newspaper, and races up the stairs two steps at a time. She delivers the final newspaper, checks her watch, exhales with relief, and smiles proudly. Outside the window, the chaotic city continues rushing by as she gives the camera a playful “Made it!” expression. Dynamic cinematic camera, fast tracking shots, low-angle bicycle shots, overhead city shots, whip pans, comedic timing, realistic motion blur, detailed city traffic, expressive character animation, warm sunrise lighting, cinematic depth of field, polished AAA animated-film quality, 4K, 24fps, 16:9. Negative Prompt: No subtitles, no on-screen text, no logos, no watermarks, no duplicate characters, no distorted anatomy, no flickering, no unnatural motion, no low-detail environments.

Athletic person stretching at sunrise, city skyline background, golden hour lighting, minimalist activewear, cinematic photography

A stressed young corporate woman leaving work late encounters a confused Viking warrior who has somehow appeared in modern-day America, then becomes his reluctant guide to fast food, shot as a cinematic fish-out-of-water buddy comedy with realistic nighttime photography and energetic multi-angle camera work. 0-4 seconds: Smooth backward tracking shot on a rainy downtown sidewalk. A 26-year-old woman in a beige trench coat, office clothes and white sneakers walks toward camera carrying a laptop bag and coffee. Woman: "I swear, if one more thing happens tonight..." 4-7 seconds: Sudden whip-pan toward an alley as a violent blue-white flash erupts. Trash bags flutter outward from the shockwave. A huge Viking warrior with braided blond hair, fur cloak, leather armor and round wooden shield stumbles into the street coughing through smoke. Viking: "MIDGARD?!" 7-11 seconds: Medium handheld reaction shot. The woman freezes. The Viking raises his axe suspiciously, then spots her coffee cup. Woman: "Please tell me you're filming a commercial." Viking: "What potion is that?" 11-15 seconds: Tight insert shot. She cautiously hands him the coffee. He takes one heroic gulp, immediately recoils from the heat, spraying coffee sideways. Viking: "THE CUP ATTACKS!" Woman bursts out laughing. 15-19 seconds: Side-mounted tracking shot from outside her small car as the enormous Viking squeezes awkwardly into the passenger seat, shield wedged against the roof. Woman: "I'm taking you somewhere civilized." Viking: "To your king?" Woman: "Wendy's." 19-23 seconds: Drive-thru POV from inside the car. Viking leans toward the speaker box. Speaker: "Welcome, what can I get started for you?" Viking, terrified: "THE WALL SPEAKS!" 23-27 seconds: Fast dolly-in inside the restaurant. Viking stares reverently at an enormous cheeseburger, slowly lifts it with both hands, takes a bite and becomes emotional. Viking: "Your people possess great wisdom." 27-30 seconds: Exterior crane pull-back. They sit on the hood of the parked car beneath glowing restaurant signage, Viking drinking a giant soda through a straw while holding his axe. Woman: "So... where exactly are you staying?" Viking: "Your couch." Additional Notes: Keep the woman’s trench coat, office clothes and appearance identical throughout. Viking always retains braided blond hair, fur cloak, armor, shield and axe. Treat the Viking and props with realistic physical weight. Modern people react naturally in the background without turning the scene into chaos. Rain remains visible throughout exterior shots. Comedy comes from sincere performances rather than exaggerated cartoon acting. Photorealistic cinematic realism, believable skin, natural motion blur, realistic reflections and fabric physics. No subtitles.

Generate a 15-second ultra-realistic cinematic close-quarters combat video, strictly 16:9 landscape, native 4K, 24fps. The overall style is a gritty neo-noir action scene on the streets of Istanbul on a rainy night. The image must look like a real live-action shoot, not AI beauty, not a digital human, not game CG. [Material Binding] image1 = First person / Protagonist reference. image2 = Second person / Attacker reference. Use image1 for facial features, proportions, hairstyle, color, skin texture, body proportions, age, clothing logic, shoes, accessories, and overall temperament reference. Use image2 for the attacker's facial features, proportions, hairstyle, color, skin texture, body proportions, age, clothing logic, shoes, accessories, and overall temperament reference. Both are clearly adult female fighters. Maintain recognizable identities, but allow for natural live expressions, wet hair swaying, breathing rise and fall, muscle tension, pain reactions, and realistic movement deformation. Do not fix faces into stiff photos. Do not make character concept art, collages, storyboards, labels, text, borders, or panels. [Real Texture Requirements] Must avoid AI beauty look. Retain natural human skin texture, slight asymmetry, real pores, fine under-eye texture, wet stray hair, skin reflections after being wet by rain, natural blinking, jaw tension, unsteady breathing, muscle exertion, and real pain reactions. The face cannot look like a plastic doll, a digital avatar, overly airbrushed, or have glass eyes or overly perfect features. [Scene] Istanbul night street after rain. The entire fight takes place only on this street. The street is narrow and real, the ground slippery and shiny, with puddles, rubble, scattered trash, parked vehicles, old building walls, roadside railings, dim streetlights, distant neon glows, damp mist, and wet reflections. The atmosphere is gritty, dangerous, cold rainy night, with a neo-noir quality. A dark parked sedan appears fixed at the right side or back of the street from the start. It must maintain the same spatial position throughout the shot, eventually becoming the target the protagonist hits after being kicked flying. Do not let the car suddenly appear. [Core Action Principles] Real close-quarters combat, not dance, not martial arts, not superpowers. Every attack must have a clear causal chain: Charging -> Striking path -> Contact or grazing -> Body reaction -> Center of gravity change. Prohibit fake hitting, reacting across space, body clipping, falling without reason, or weightless flying. Actions must be fast, fierce, and clear, but not chaotic. Both must look like real trained combat actors. The attacker is more active and fierce; the protagonist starts calm, contemptuous, and precise, but is finally hit by a heavy blow and knocked back, in a losing state. [Photography] A continuous handheld long take, no cuts, no black screen transitions, no teleportation. The camera path must be clear: Protagonist's shoulder OTS opening -> Natural move to side profile close-up -> Close-quarters combat orbiting tracking -> Following the direction the protagonist is kicked -> Pushing into the protagonist's face on the damaged car. Handheld photography should be close, fluid, and immersive, but not so shaky that actions cannot be seen clearly. Use brief speed changes only at key close contact and final heavy impact moments; do not use slow motion for the entire segment. [Timeline] 0.0–2.5s Starts with an OTS shot behind the protagonist's shoulder. The protagonist stands in the foreground of the slippery street, back and shoulders occupying part of the frame. The distant attacker rushes quickly from the other end of the rainy night street, feet stepping through puddles, splashing water. The camera moves past the protagonist's shoulder, seeing the parked dark car at the right or back of the street. Sound of rain, footsteps, distant car sounds, and low ambient noise are present. The protagonist is not flustered, only slightly turning her head, eyes calm. 2.5–5.5s The camera naturally orbits to a side profile angle of the two. The attacker rushes to close range, throwing a fast jab or right punch at the protagonist's face. Use a very short speed change only at the moment the punch grazes the side of the face. The protagonist does not dodge significantly, does not lean back, does not retreat. She only does a minimal head retreat and jaw turn at the last moment. The fist must clearly graze a few centimeters in front of her cheek; the fist, cheek, and gap between them can be seen. The wind of the punch slightly moves her hair and collar. The protagonist's mouth shows a very slight contemptuous smile, as if thinking the opponent is too slow. The expression should be natural, not exaggerated. 5.5–8.5s Speed returns to normal. The attacker immediately continues to press, throwing a second blow and close-range combination. Sequence of actions clear: Attacker left hook -> Protagonist parries with forearm -> Attacker low knee or short kick -> Protagonist resolves with a small step back with the rear foot -> Protagonist counters with a short elbow or punch to the attacker's chest or shoulder line. Every movement must have a sense of contact. Blocking must actually change the attack path. The side hit must react according to the direction of force in shoulders, head, chest, and footsteps. Footsteps on wet ground will slip, but not lose control. 8.5–11.5s Close-quarters combat escalates. The attacker is more fierce, pressing continuously. The protagonist attempts to control the attacker's arm and throws a quick counter. Includes: Low angle seeing the attacker's preparation for a spin kick or strong side kick; Protagonist first dodges a dangerous close attack; Then the two briefly clash, with wrist control, shoulder bumps, and elbow parries. The camera orbits half a circle around them, keeping faces, hands, feet, and ground reactions clear. The protagonist still seems to have the advantage, but the attacker is not suppressed; she is waiting for a heavy blow opportunity. 11.5–13.0s The attacker suddenly explodes, throwing a powerful spinning back kick targeting the protagonist's chest or upper center. Use brief slow motion and an elongated low-frequency impact sound at this moment. The attacker's heel or sole clearly hits the protagonist's upper body, not a miss. The protagonist's body is driven by the impact starting from the chest, shoulders leaning back, feet leaving the wet ground, water and dust splashing from the floor. The impact must be realistically weighted, not like flying away weightlessly. 13.0–15.0s The protagonist is kicked flying toward the dark parked sedan that appeared at the start. The camera quickly follows her backward direction. Her back and shoulders hit the car body and front windshield area; the windshield shatters, glass fragments scatter, but not bloody. She falls heavily onto the car body or broken windshield, her body collapsing as if losing power. The attacker stands on the wet street, maintaining a strong fighting stance, looking at her coldly like a victor. In the last second, the camera no longer emphasizes the attacker but slowly pushes from the attacker's direction to the protagonist's face lying on the damaged car. The protagonist's face has rain, fatigue, pain, and resentment, still maintaining image1 identity features. The attacker remains slightly blurred in the background. The end should not cut to black, stopping on the protagonist's broken but conscious expression.

CAMERA: Handheld DV 16mm daily vlog footage. The video MUST begin with her holding the camera at arm's length in selfie mode while stepping out of an old stone doorway into a narrow street. The first 20–30 seconds are entirely handheld. Later she occasionally sets the camera on a windowsill, café ledge, or low seawall for wider shots. Keep subtle handheld shake, drifting composition, autofocus hunting, rushed reframing, uneven zooms, exposure breathing, brief accidental face cropping, and imperfect framing throughout. The camera itself is never visible. LOOK: Warm analog tape texture with gentle film grain, softened sharpness, subtle halation where sunlight cuts between buildings, faint salt-air haze, realistic skin tones, low contrast, tiny exposure shifts, natural motion blur. Authentic everyday footage, not a polished commercial. STYLE: A quiet Mediterranean morning coffee-run vlog. Unhurried, warm, spontaneous. She talks to the camera in short sentences with comfortable pauses, reacts to the street around her, laughs easily. CHARACTER: ELENA — a Mediterranean woman in her mid-20s. Dark wavy hair in a loose bun, olive skin, minimal makeup, small gold hoops, oversized white linen shirt, denim shorts, leather sandals, canvas tote on one shoulder. SETTING: A sun-washed coastal old town just after sunrise. Narrow limestone alleys, pastel shutters, bougainvillea spilling over walls, glimpses of the sea between buildings, distant church bells and scooters, almost no people. A sleepy stray cat on a warm step. SCENES: The vlog opens in selfie mode as Elena steps out of the doorway into the bright alley, squinting. "Morning." She walks, sunlight strobing between buildings across her face. "Okay it's already hot. Of course it is." She reaches a tiny corner kiosk and orders an iced coffee, camera still handheld, framing half-wrong. She sets the camera on the counter ledge for a wide shot while she waits, drumming her fingers. A stray cat appears at her feet and stares up. She crouches down to it. "Hi. You again." The cat rubs against her leg. She laughs. "I have a bodyguard now." She picks the camera back up, coffee in the other hand, and walks toward the water, the cat trotting behind for a few steps. She sits on the low seawall, sea behind her, and takes the first sip. "Worth it. Every time." She looks off at the water for a moment, hair moving in the breeze, then back to the lens. "Okay. Work now." A small wave. "See you later."

Create a fast-paced cinematic 15-second vertical 9:16 superhero edit blending photorealistic live action with an original comic-book animation aesthetic. 0–4 SEC — CITY RUN A young woman with long brown hair, oversized light-pink shirt, white tank top, light-blue jeans, and white sneakers runs through a modern European-inspired city. She films herself with a pink smartphone, smiling at the camera. Smooth tracking shot suddenly transitions into bold comic-book linework, halftone textures, vibrant colors, and speed lines. 4–8 SEC — HERO ACTION She jumps between rooftops as two completely original masked web-slinging heroes appear. One wears a white, charcoal, and teal suit; the other wears crimson and midnight blue. No recognizable existing superhero designs. Energy webs shoot from their gloves as they swing between skyscrapers. 8–12 SEC — COMIC CHAOS First-person web-swinging POV through a futuristic city at golden hour. Neon signs, floating debris, graphic panels, motion blur, and an original “WHOOSH!” graphic create explosive comic energy. She lands dramatically on a rooftop. 12–15 SEC — FINAL HERO SHOT She crouches on a rooftop with her hair blowing in the wind as the camera pulls back to reveal a vast fictional city skyline glowing pink and orange. Final text: WEBRUNNER RISE. SWING. RUN. STYLE: Premium cinematic live action, original graphic-novel animation, bold outlines, halftone textures, vibrant colors, realistic hair and fabric physics, dynamic camera movement, dramatic lighting, seamless transitions, photorealistic detail, 4K, vertical 9:16. AVOID: Existing superhero characters, recognizable costumes, logos, franchise imagery, real-world landmarks, copied animation styles, distorted faces, warped hands, inconsistent characters, random text, cheap CGI.

Photorealistic fixed surveillance camera (CCTV) footage. A traffic surveillance camera mounted on a high pole captures a wide-angle view looking down slightly at a two-lane road and the sidewalk beside it. Natural daylight, faint contrast characteristic of real security cameras, slight sensor noise and light wide-angle distortion, with a date and time timestamp in the corner of the screen. The camera remains completely fixed and stationary throughout. [0s-9s] A peaceful daily scene. A father is walking slowly on the sidewalk holding hands with a young child. Several cars pass by on the road at normal speeds. [9s-12s] A passenger car suddenly loses control, tires screeching as it crosses the lanes and charges straight toward the parent and child on the sidewalk. [12s-15s] At the final moment, the father snaps around and reflexively scoops up the child, jumping aside. The car plows through the exact spot where they stood, stopping in a cloud of dust. Style: Photorealistic quality of actual surveillance recordings. Fixed wide shot, no cinematic camerawork. Ambient sounds only.

The downtown DMV headquarters in Sacramento, California, security guards are patrolling the parking lots for state government employees. Three male blonde security guards in their late 20s to early 30s are on bike patrol and ride to the back parking...

A glowing neon origami crane flying through a dark, rainy cyberpunk alleyway, illuminated by flickering holo billboards, fast cinematic blur, hyper-detailed, 8k ultra-realistic

A 20 second desert caravan journey, glowing sun, rolling dunes, a mirage resolving into a golden city.

Create a 30-second ultra-realistic boyfriend-vlog video filmed as one continuous, unbroken handheld take. Use the woman from the reference image as the main character and preserve her exact facial identity, hairstyle, facial features, body proportions, and overall appearance throughout the entire video. She must remain clearly the same person in every moment. The footage should feel like her boyfriend casually filming her during an ordinary day in Tokyo with a small mirrorless camera or smartphone. Natural handheld movement, slight camera shake, imperfect framing, realistic autofocus, subtle exposure changes, occasional motion blur, and spontaneous reactions. She never poses or performs for the camera; she behaves naturally and sometimes forgets she is being filmed. 0–6s — Morning apartment. Soft daylight enters a small Tokyo apartment. She sits near the bed fixing her hair and getting ready. She notices the camera, smiles, laughs, and playfully gestures for her boyfriend to stop filming. Keep the moment intimate and spontaneous. 6–12s — Tokyo neighborhood. She walks ahead carrying a small bag while the camera follows behind. Quiet residential streets, bicycles, pedestrians, small shops opening, natural morning activity. She occasionally looks back at the camera and casually interacts with the person filming. 12–18s — Local food stop. They enter a cozy neighborhood restaurant. She sits down and tries ramen or another local Japanese dish. Capture close handheld moments of her picking up chopsticks, tasting the food, reacting naturally, and laughing when it is hotter than expected. The boyfriend reacts naturally from behind the camera. 18–24s — Afternoon exploration. She walks through a lively Tokyo neighborhood, browses small shops, looks at interesting objects, takes photos, and occasionally glances back at the camera. The camera naturally shifts between her face, hands, surroundings, pedestrians, storefronts, and small details of everyday life. 24–30s — Evening ending. Tokyo streets are illuminated as night falls. She walks slightly ahead, then turns back and smiles at the camera. Cut naturally to them riding a train home; she sits beside the window watching Tokyo lights pass outside. The camera gently moves closer as she rests quietly, ending like an authentic personal memory. VISUAL STYLE: Ultra-realistic candid boyfriend travel vlog, documentary realism, natural lighting, authentic human expressions, imperfect handheld camera behavior, realistic skin texture, natural body movement, subtle motion blur, believable Tokyo atmosphere. No commercial/ad aesthetic, no dramatic posing, no cinematic perfection, no artificial transitions, no CGI appearance, no text, no logos, no face changes, no identity drift, no wardrobe changes, and no character inconsistency.

A cinematic superhero-inspired short film. A young man wearing a sleek red-and-black futuristic suit stands alone on a rainy rooftop at sunset, overlooking a neon-lit cyberpunk city. Bright orange holographic energy (#FF4900) pulses through the suit's glowing lines. Suddenly, a massive dimensional portal tears open across the sky as robotic alien drones invade the city. The hero leaps from the rooftop in slow motion, soaring between skyscrapers with fluid aerial movements. The camera follows with dynamic drone shots, close-ups, and dramatic orbiting angles. He dodges laser blasts, performs high-speed flips, and unleashes brilliant orange energy attacks that destroy multiple drones. Explosions illuminate the skyline while civilians escape below. The action seamlessly transitions into an intense mid-air battle with cinematic lighting, volumetric fog, realistic reflections, debris physics, and ultra-detailed VFX. The final scene shows the hero hovering above the city as the portal collapses, revealing a peaceful sunrise. A glowing orange hologram appears reading "Created with Seedance 2.5" before fading into the skyline. Ultra-realistic, Hollywood blockbuster quality, photorealistic, 4K, smooth camera motion, consistent character appearance, dramatic storytelling, emotional pacing, film-grade color grading, epic orchestral soundtrack feel, designed as one continuous 30-second AI-generated cinematic sequence.

Montage, multi-shot live-action action movie. Photorealistic, high-contrast teal and orange grading, 35mm film look. [Image 1 defines the character's appearance (front), Image 2 defines the back. Used for appearance reference only, ignoring green backgrounds. Fully maintain reference image appearance.] The masked attackers appearing throughout the video are exactly 4 people, split into two waves: Wave 1 is 2 people coming out of an alley, Wave 2 is 1 person jumping out of a parked van and 1 person jumping over a car hood. No other characters appear. Each attacker is defeated only once and remains down on the wet asphalt after falling. Shot 1 (0-3s): Neon night city in the rain. Reflections shimmer on wet asphalt, steam rises from vents. A man in a blue check suit and helmet walks quietly on an empty sidewalk, rain flowing down the helmet. In front, 2 masked attackers from Wave 1 emerge from a dark alley and block the path; he stops. Slow tracking shot from behind. <Heavy rain, distant thunder> Shot 2 (3-6s): The 2 from Wave 1 attack. He slams the first into a metal shutter, sparks fly in the rain, and the first collapses and remains down. Handheld tracking. <Heavy impact sounds, metal sounds, rain> Shot 3 (6-10s): Slow motion. He leans back under a baton swung by the second person; raindrops stand still in the neon light. Returns to normal speed with a speed ramp, slams that same attacker hard onto the wet asphalt with a splash; the second remains down. Now the 2 from Wave 1 are immobile. Low angle. <Deep whoosh, heavy impact, splash> Shot 4 (10-12s): The sliding door of a parked van bursts open and a 3rd person jumps out. Immediately after, a 4th person jumps over the car hood in the backlight of the headlights. Rain blows through the light beams. Medium wide, slight handheld shake. <Van door sound, footsteps on wet ground> Shot 5 (12-15s): The 4th person throws a punch. He dodges sideways and sinks the 4th person into the wet asphalt with a sharp kick; the 4th remains down. Handheld tracking. <Sharp impact sounds, rain> Shot 6 (15-17s): The 3rd person from the van tackles. He turns to dodge and crushes them along with wooden crates stacked by the curb. Broken wood fragments fly in the rain; the 3rd person also remains down. Now all 4 people are down and immobile on the wet street. Handheld close-up tracking. <Wood breaking sound, heavy fall sound> Shot 7 (17-20s): Medium close-up in the pouring rain. Neon reflections tremble in puddles, the 4 fallen people stay motionless in the background. As he fixes his red tie, a bright yellow 8-bit style pixel smile lights up on a black screen, glowing warmly through the rain. Slow dolly-in, low angle. <Rain sound, soft retro beep> Maintain the same helmet, blue check suit, red tie, olive backpack, and black leather shoes across all shots. Sharp with fixed focus, stable image, no deformation. No music, no on-screen text except the pixel smile on the helmet screen.

Create a 10-second ultra-realistic cinematic lifestyle video featuring the same stylish young woman exploring Birmingham. 0–2 sec: She walks through a vibrant urban street and notices a large colourful street-art mural. 2–5 sec: She approaches the mural and is handed a spray-paint can by a local artist. She adds a small artistic detail to the wall. 5–7 sec: Close-up of her laughing and stepping back to admire the artwork, with paint marks visible on her gloves. 7–10 sec: Camera pulls back to reveal the full colourful mural and the woman standing confidently beside it, posing for the camera. Style: Photorealistic, cinematic urban lifestyle aesthetic, authentic Birmingham atmosphere, vibrant street art, natural facial expressions, smooth camera movement, realistic lighting, consistent AI-model identity, 4K, dynamic Instagram/Reels style.

Generate a 15–20 second, 9:16 vertical, strictly one-take ultra-realistic cinematic video. The entire shot features only one continuous aerial follow-cam, with no cuts, transitions, jump cuts, or resets. [Reference Image Mapping] Image 1: Action and close-up composition for landing on a car roof. Image 2: Action reference for parkour across moving car roofs. Image 3: Sunset sky, NYC brick neighborhood, web-swinging, and final wide-angle atmosphere. Image 4: First frame, character identity, face, hairstyle, clothing, and bedroom environment. [Character Consistency] The character is always the young Asian woman from Image 4: same features, medium-length straight black hair with bangs, off-white short-sleeved knit top, off-white high-waisted shorts, and white sneakers. No face swapping, no costume changes, no superhero suits, and no masks. Body proportions remain stable and natural throughout. [Action Timeline] 0–3s: Camera starts inside a warm, cluttered NYC bedroom. The girl sits cross-legged on a wide windowsill. Golden sunset fills the room. She looks at the camera, then turns toward the city. The camera moves past her side smoothly toward the window. 3–5s: She stands up, shoots a thin white web from her wrist to a nearby building, and jumps out the window. The camera follows her through the window without a cut. 5–8s: She swings down toward the street with realistic gravity and wind effect. She releases the web and lands in a crouch on a moving black car roof. The roof slightly compresses under the impact; she absorbs the shock with her arms and legs—no floating. 8–12s: She sprints forward across car roofs. The camera follows closely at a low angle with a slight Dutch tilt. She performs three distinct parkour jumps between roofs with clear landing feedback. 12–15s: Near the end of the block, she jumps and shoots a new web at a red brick building. The rope stretches and rebounds naturally as she swings between buildings. 15–20s: The camera follows the second swing, pulling back into a super wide-angle shot showing the NYC skyline and brick apartments under the sunset. The character swings toward the sun, becoming a small figure in the center. Ends with a stable Dutch angle. [Physics and Camera Requirements] Realistic gravity, inertia, and impact mechanics. Webs must connect wrist to fixed anchors. Camera moves like a high-agility FPV drone—smooth, fast, and stabilized with natural motion blur. Super-realistic film texture, detailed skin and hair, cinematic color grading, 8K quality. [Strict Avoidance] No cuts, no costume changes, no superhero uniforms, no floating or ragdoll physics, and no sudden scene changes.

Create a 15-second ultra-realistic vertical lifestyle video using the reference image. Keep the exact same woman, facial identity, hairstyle, outfit, body proportions, and appearance throughout. Story: A rainy afternoon at a modern train station. She walks toward the platform and notices her train leaving, slightly quickening her pace. She arrives just as the doors close and watches it depart, mildly disappointed. She exhales, gives a tiny amused smile, and checks the departure board. She discovers the next train arrives in a few minutes, relaxes, gives the camera a playful subtle shrug, then sits on a nearby bench and calmly checks her phone. Shots: Natural cinematic cuts: rear three-quarter walking shot → side platform shot → front close-up → over-the-shoulder departure board → final front tracking shot. Performance: Understated, believable acting. Emotional progression: focused → realization → mild disappointment → amusement → relief. Natural blinking, breathing, eye movement, walking, and body language. Environment: Modern clean station, rainy gray daylight through large windows, wet floor reflections, distant passengers, realistic moving train, soft station ambience. Camera & style: Authentic handheld smartphone footage, subtle camera movement, natural autofocus, realistic depth of field, photorealistic Korean lifestyle vlog aesthetic, cinematic but candid, natural skin texture and colors, 24fps. Avoid: Face/identity changes, outfit changes, exaggerated acting or running, distorted hands, extra fingers, unnatural walking, unrealistic train movement, floating objects, sudden lighting/environment changes, excessive camera shake.

High-quality anime footage. High-quality anime, 2D hand-drawn. Facial expression changes proceed in the order of eyes -> eyebrows -> eyelids -> mouth -> gaze. [Character Fixation] Maintain the character from Image1 as the same person throughout. Image1 is the only person appearing on screen. Small and delicate round facial features, large purple to lilac jewel-like eyes, multiple white catchlights, thick upper eyelashes, faint cheek redness, long pastel pink hair tied in two bunches on the sides, loose curls on the forehead, purple flower ornaments and dark blue ribbons. White short-sleeved sailor collar top, dark blue and gold trim, lavender pleated skirt, dark blue frill hem, and purple jewel belt do not change. Only expressions, gaze, natural mouth movements, posture, breathing, and the swaying of hair and skirt in the wind may change. Do not produce different people, facial averaging, hairstyle changes, outfit swaps, doubles, mirror images, reflections, other people, hands, shadows, or silhouettes. The conversation partner does not appear even off-screen; the presence of a date is expressed only by her gaze toward the diagonal left front. The camera is objectively third-person throughout. [Art Style Fixation] Maintain the delicate hand-drawn line art and pale purple-tinged line colors, transparent multi-layered shadows, soft gradients, milky pink and lavender, dark blue and gold accents, fine hair textures, jewel-like eye sparkle, elegant highlights, and detail in the skirt pleats and flower ornaments of Image1. The summer town, cafe, riverside, fireworks, and sky are all drawn as the same high-density 2D hand-drawn background art. Do not use thick lines, heavy outlines, simplified TV anime, flat cell shading, low-density backgrounds, smooth CG/3D conversion, photo-like textures, dull colors, expressionless faces, or mixed styles. The reference image is only for appearance and style; it does not reproduce the stone corridor, castle, pose, composition, or camera distance of the original image. Every cut is a new summer urban date scene. [Prop and Audio Fixation] There is only one transparent iced tea cup, owned only by Image1. At the beginning of Cut 1, Image1 holds it in both hands in front of her chest, and after taking a sip, Image1 herself quietly places it on the cafe table with her right hand at the end of Cut 1. The cup remains on that table and is not shown from Cut 2 onwards. A refreshing romance game-style instrumental BGM without vocals. Use soft electric piano, clean guitar, modest shakers, and pale synth pads, naturally building excitement from evening to night. Synchronize small sounds like ice clinking in the cup, distant city ambient sounds, rustling clothes, footsteps, and the sound of fireworks launching with visible actions. Lower the BGM slightly just before and during dialogue to make Image1's Japanese voice clearly audible. Do not include singing, narration, the other person's response, or other voices. Cut 1: 0.0–3.6s | Street corner cafe at dusk. Camera position is three-quarters to the left front of Image1, at eye level, character size is chest-up, background vanishing axis extends from the back of the table to the back right of the street illuminated by the sunset. The goal is to simultaneously read the sip of iced tea and the first flicker of excitement in Image1's eyes. Image1 holds the transparent cup of iced tea in both hands and takes a sip while smiling. The ice makes a clear sound once, and the sunset softly illuminates the rim of the cup and the outer edge of her hair. Image1 firmly places the cup on the table with her right hand, shifts her gaze slightly to the diagonal left front, and her eyes and the corners of her mouth brighten slightly. The camera slides sideways for a short distance at the same height, and the shadow of the cafe awning momentarily crosses in front of the lens to connect to the next street. Cut 2: 3.6–7.2s | Shopping street at sunset. Camera position is three-quarters to the left front of Image1, slightly higher than the waist, character size is medium shot, background vanishing axis continues to the far left of the shopping street in her direction of travel. The goal is to read the excitement in her steps and words from her face. Image1 naturally takes a step forward. As a short preliminary action before speaking, her eyes brighten slightly, eyebrows lift just a bit, eyelids soften and narrow, the corners of her mouth relax, and finally she looks toward the diagonal left front. Image1 herself naturally speaks once: "Today is really fun." Completely synchronize lip movements, breath, expression, and voice. During the speech, lower the BGM to bring the voice forward, and in the short lingering sound at the end, the clean guitar and shaker return. A foreground blur of street tree leaves briefly crosses the lens, switching to the night riverside. Cut 3: 7.2–11.0s | Riverside promenade changed to blue-purple night. Camera position is the right side of Image1, at eye level, character size is a medium shot where the profile can be read, background vanishing axis extends from the riverside promenade to the sky of distant fireworks. The goal is to show the moment the change in the night sky reaches Image1's surprise and excitement. Image1 looks up at the sky once. The moment her hair bunches and skirt hem sway slightly late in the night breeze, a large purple, peach, and gold firework bursts in the distant night sky. Do not overlay firework light as dots or lines on the face; return it only as soft surface light to the outer edge of the hair and the edges of the outfit. At the moment the firework bursts, add warm synth pads and thin strings to the BGM to expand it by one stage. The camera leans in slightly from the right, maintaining a distance where the eyes and smile changing from surprise to excitement can be read. Cut 4: 11.0–15.0s | Afterglow of fireworks and title. Camera position is three-quarters to the left front of Image1, at eye level, character size is chest-up, background vanishing axis opens to the back of the night sky on the right. The goal is to let Image1's face and the title be read together as a single afterglow without interfering with each other. Keep Image1 in a chest-up size on the left third of the screen, leaving only the right night sky open as white space for the title. Image1 continues to breathe naturally without stopping with a clear smile, blinking slowly once while looking at the fireworks, shifting her expression slightly from the profile to the diagonal front. On the right side of the night sky, display a clean two-line title treatment well away from the face, eyes, mouth, and hair ornaments. In small milky white characters: "This summer, my encounter with you turns into excitement," and below it in large deep indigo with pale lavender edge light: "Midsummer Night." Display it as a work title rather than a subtitle, and do not show other readable text, logos, or watermarks. The camera pulls back slightly without making Image1 too small. At exactly 15.0 seconds, both Image1's face and the title are completely readable, and cut with a short piano afterglow just before the last firework tail disappears in the night sky.

Hollywood california , labor hall , labor ready, two young men one in 30s one early 20s are talking and another young man walks buy , he is limping and holding his ribs, he recognizes the man in his 20s and they talk, tge 20 yr old asks what happened