LogoAhaPrompt
  • Prompts
  • Featured
  • Models
  • Submit
  • Blog
AhaPromptAhaPrompt

AI Prompts for Images, Video & Audio — Every Model, One Library

Built withLogoNEXTY.DEV
Explore
  • All Prompts
  • Featured Prompts
  • Image Prompts
  • Video Prompts
  • Audio Prompts
  • AI Models
  • Submit a Prompt
  • Blog
Languages
  • English
  • 中文
  • 日本語
  • 한국어

Copyright © 2026 AhaPrompt All rights reserved.

Privacy PolicyTerms of Service
HomeAI Prompt LibraryAI Video PromptsArmored Warrior vs Courtyard Ghouls

Armored Warrior vs Courtyard Ghouls

ByteDanceSeedance 2.5ByteDancevideo

A mature adult heavy warrior is the sole protagonist: powerful build, fully enclosed in original blue-black gunmetal plate armor, face永远不可见; helmet is smooth enclosed type, rounded top, only a narrow horizontal eye slit, no emblems, no heraldry, no familiar franchise elements. Pauldrons wide and round, elbow armor, knee armor, steel gauntlets, segmented waist armor clear and stable, joints reveal only dark chainmail and thick padding. He always uses the same two-handed dark steel longsword, simple straight crossguard, size, material, and grip orientation always consistent, cannot disappear, duplicate, bend, or switch hands without cause. Enemies are original non-human courtyard deformed gaunt fiends: the two active gaunt fiends must be distinguishable from each other, both hairless, gray-white skin, abnormally tall and emaciated, ribcage exposed, hunched spine, long arms and long fingers, but must be anatomically coherent, each with only two arms, two legs, five fingers. The final giant brute is a separate individual, about one point five times the warrior's mass, broader shoulders and back, thicker limbs, muscular hunchbacked torso, more pronounced spinal bone ridge, absolutely must not become a clone or twin of the gaunt fiends. All characters completely independent, identity, body proportions, silhouette, limb relationships always clear, absolutely no fusion, absolutely no shared limbs, absolutely no facial drift, absolutely no limb deformation. Style: high-end fully computer-generated AAA dark fantasy action game cinematic cutscene, original high-end real-time engine rendering, stylized realism, must clearly be polished game CG, not live-action photography, not 2.5D, not animated series, not illustration. Physical materials are blue-black gunmetal armor, cold steel blade, sculptural creature muscle, full three-dimensional parallax, stable environmental geometry, cinematic procedural fog, controlled game cinematic motion blur, refined character animation. Native 16:9 aspect ratio, 24fps, full use of the entire frame, no vertical screen, no blurred duplicate borders, no black bars, no subtitles, no text, no logo. Scene is an original abandoned stone fortress courtyard at night: uneven dark cobblestones, broken arched windows, weathered masonry, recessed doorways, scattered pale corpses, dense blue-gray ground fog slowly flowing low along the ground, no modern objects. Color locked to cold moonlight cyan-blue, deep navy shadows, low-saturation stone and gray-white skin, restrained silver-blue highlights on armor edges and sword steel, near-black dark blood is the only warm-leaning visual accent; low exposure but readable silhouettes, volumetric depth, soft anamorphic widescreen highlights, slight halation and fine grain. Camera rules: one continuous shot, no cuts, no jump cuts, no true hard cuts, duration strictly about 20.00 seconds. Use ARRI Alexa 35 with 35mm Cooke anamorphic virtual cinematic lens language, but maintain game engine texture. Camera low angle, medium shot to medium-long shot combat composition, close enough to clearly read hands, feet, weapons, and contact. Use controlled virtual dolly tracking and restrained orbital turns, not exaggerated handheld; shallow to medium depth of field, warrior and current attacker clear, distant ruins fade into fog. Forbidden hero angles, forbidden dramatic frontal low-angle shots, forbidden facial close-ups, forbidden showy crane or drone movements, forbidden additional editing. Must be unmistakable ultra slow motion throughout: all attacks like extremely high-speed cinematic camera capture then slowed down, movements have smooth intermediate poses, real inertia, resistance, delayed reactions, and weight; no normal-speed bursts, no freeze, no looping motion, no random extra attacks. 0.00-2.75 seconds: Camera starts from a slightly low frontal medium shot, warrior near frame center, feet planted wide and firm on wet cold cobblestones, body slightly facing screen right, helmet tilted slightly down as if listening. Both hands grip sword near waist, blade angled toward upper right of frame. Pauldrons show only slight breathing rise and fall, steel fingers slightly tighten. One gaunt fiend lurks low behind to the left rear of screen, another different gaunt fiend gradually emerges from fog on screen right, shoulders slumped forward, long arms hanging down. Camera drifts almost imperceptibly to the side, fog slides past behind legs, cold light slowly crawls along curved armor surfaces, warrior's center of gravity makes only tiny shifts. 2.75-4.00 seconds: The right gaunt fiend suddenly makes an elongated horizontal lunge, rear foot drives, torso almost parallel to ground as it rushes toward the warrior's helmet, both arms reaching toward screen left, long fingers spread. Warrior senses it and lowers his center of gravity, hips and shoulders turn toward the incoming body, both hands raise the sword to a position above and slightly behind the helmet forming a high guard, elbows folded, wrist line aligned, blade nearly horizontal. He does not dodge away, but stays beneath the creature's forearms, camera only gently pans to keep both bodies in frame. 4.00-5.90 seconds: The gaunt fiend's hands sweep past above and in front of the helmet, torso crosses from screen right to left near the foreground. Warrior drives from his rear foot planted firm, waist and shoulders rotate in sync, completing a single, clean two-handed horizontal counter slash beneath the incoming long arms, must clearly show wind-up, contact, cut-through, and completion. The blade passes through the waist only once, finally stopping at chest height and pointing toward screen right. Only after the sword has passed do the upper body and lower body separate cleanly. Blood extremely restrained, dark near-black: a small amount of suspended droplets and thin mist trailing along the sword line, the lower body still sliding forward from its original momentum. 5.90-7.87 seconds: The severed upper body tumbles into the lower left foreground, shoulder sinking and rotating first, then heavily pressing onto the stone surface, both arms dragging and folding beneath it. Warrior maintains balance, follows the sword's inertia to decelerate, helmet and upper body turn toward screen right, weight shifts to front foot, pulling the sword back to a two-handed ready position. The second gaunt fiend advances more clearly in the fog behind right. Camera does not cut, but lets the falling upper body briefly brush past the near foreground creating a natural occlusion, using this physical occlusion to continuously transition into a three-quarter rear view, simulating the composition of what would originally be a second shot, but absolutely no hard cut. 7.87-9.25 seconds: The same continuous shot is now in a medium-long three-quarter rear angle, warrior's armor undamaged, still the same longsword. The second gaunt fiend extends one long arm from screen left grabbing toward the helmet. Warrior turns, one armored forearm raises to defend, one steel hand still controls the longsword low at hip side, blade may briefly near frame edge but cannot disappear. He presses weight onto support leg, near-camera-side knee sharply raises to chest, then delivers a powerful forward kick into the gaunt fiend's lower chest and abdomen. Must linger on the moment of contact: boot sole compresses torso, spine bends backward, shoulders lag, both arms flung behind by inertia. Then he retracts knee and plants foot steadily, no sliding. 9.25-11.30 seconds: The gaunt fiend staggers trying to recover balance, warrior re-establishes a clear two-handed sword grip. He sinks shoulders, rotates hips, pulls the sword up from behind upper right shoulder, both hands raising it above the helmet. After the preparation completes, he delivers a downward diagonal chop, cutting through the gaunt fiend's neck and upper shoulder. The sword makes only one readable contact and continues in controlled follow-through. The head remains attached, but snaps down heavily toward the wounded side, a deep dark wound opens at neck and shoulder, restrained dark blood arc and a few suspended droplets follow the motion. Warrior ends with sword low and horizontal in front of body, both wrists still locked. 11.30-13.00 seconds: Warrior lowers sword tip, eye slit briefly downward. The wounded gaunt fiend behind has uneven shoulders, weakening knees, head hanging crooked, then begins to collapse. In the central background, the giant brute emerges from the fog, leaning forward with open mouth, both thick arms spread as it accelerates into a charge. Camera restrainedly rotates to face the threat, while keeping the wounded gaunt fiend's swaying fall in the fog, must not let background corpses suddenly spawn or move. 13.00-15.90 seconds: The brute charges into the foreground, broad chest and spinal ridge rapidly filling the frame, body creating natural foreground occlusion and physical wipe, but no edit. Warrior bends knees, lowers helmet and shoulders, ducks under the brute's sweeping arm, sidesteps diagonally out of the way. The brute's momentum carries it past, both arms continuing to swing forward, shoulders not yet able to recover. Camera follows the interweaving bodies. As the brute leans forward off balance, warrior steps close around to its rear, both boots planted firm, both hands return to the sword hilt, longsword raised vertically above helmet and right shoulder, spine straightened, both elbows loaded preparing for a downward chop; brute's back muscles tense, center of gravity still plunging forward. 15.90-18.85 seconds: Warrior drives from rear foot, sinks shoulder, delivers a heavy vertical chop down along the brute's upper back and spine. Must show blade descent, contact, and full follow-through as the same strike. A long narrow dark wound opens from upper back down along the centerline of the spinal ridge, only one restrained near-black blood line. Brute reacts gradually not instantly: shoulders hunch, elbows soften, torso folds down, knees bend, then both knees slam onto the cobblestones. Its weight continues pressing forward onto hands and forearms, chest heavily landing again. Each ground contact pushes out a ring of low fog. Warrior withdraws sword, controlled half-step avoiding the collapsing mass, lets sword hang at leg side. 18.85-20.00 seconds: Brute completes its collapse and goes still, chest and forearms pressed against stone surface. Warrior stands behind it slightly toward screen left, three-quarter rear view, boots stable, sword hanging down, helmet eye slit facing the corpse. The earlier fallen gaunt fiend and scattered pale corpses visible through fog but motionless. Ending is a restrained rather than heroic static aftermath image: brute's final shoulder residual movement stops, warrior's armor plates complete tiny shifts, a few dark blood drops fall, ground fog continues flowing across. Ends exactly at 20.00 seconds, no fade out, no black frame. Sound: completely silent output. No background music, no score, no dialogue, no breathing sounds, no ambient atmosphere, no footsteps, no weapon sounds, no impact sounds, no monster sounds, no any on-site sound effects; all sound left empty for later addition. Mandatory restrictions: strictly maintain native 16:9 full frame; 20.00 seconds; 24fps; high-end AAA game engine dark fantasy CG; ultra slow motion throughout; only one armored warrior, two distinguishable active gaunt fiends, one separate giant brute, others can only be already-fallen background corpses; same set of armor, same helmet, same two-handed longsword continuous throughout; no character spawning, no duplication, no deformation, no identity swapping, no extra attacks, no extra cuts; no modern objects; no text, no logo, no subtitles; no hero angles, no facial close-ups; no floating weapons, no rubber swords, no sliding feet, no impossible balance; dark near-black blood only in small amounts as described; all contact must have real anticipation, compression, force transfer, follow-through, momentum delay, and environmental consequences.

Details
Type
AI Video Prompts
Aspect Ratio
16:9
Duration
0:20
Updated
Oct 4, 2026
Category
CinematicFantasyDark
Source

How to use this prompt

  1. 1Copy the full prompt with the copy button above.
  2. 2Open Seedance 2.5 by ByteDance, or your preferred AI video generator.
  3. 3Paste the prompt and keep the camera and motion terms — they steer the cinematography.
  4. 4Match the example's settings (16:9 · 0:20 · Cinematic · Fantasy · Dark), then iterate — small wording changes shift the output a lot.

Frequently asked questions

Which AI model generated this video?

This clip was generated with Seedance 2.5 by ByteDance. The prompt on this page is the exact text used to create it.

What settings does this prompt use?

The example was generated with Seedance 2.5 using: 16:9 · 0:20 · Cinematic · Fantasy · Dark. Reusing these settings gives the closest match.

How do I get similar results?

Copy the prompt as-is first, then change the subject while keeping the camera, motion, and lighting terms. Video models are sensitive to wording, so expect a few runs.

Can I modify this prompt?

Yes. Prompts work best as starting points — swap scenes, adjust pacing terms, or change the aspect ratio to fit your project.

More Seedance 2.5 prompts

0:30
ByteDanceSeedance 2.5
STYLE: Hollywood disaster-action film, photorealistic live-action, ARRI Alexa 35, anamorphic cinematic lens, realistic Malaysian metropolitan city, practical stunt realism, believable destruction and debris physics, aggressive handheld tracking, natural lighting, cinematic contrast, motion blur, dust, shattered glass, sparks, immersive disaster sound. MAIN CHARACTER: Use Image 1. Maintain exact facial identity, hairstyle, clothing and continuity throughout. CORE RULE: ONE CONTINUOUS SURVIVAL SHOT. No teleportation or unexplained cuts. Every movement logically connects. The man constantly runs, stumbles, recovers, jumps, slides, climbs and changes direction. The environment is the antagonist. 00:00–00:03- EARTHQUAKE Low-angle tracking behind his feet as he runs through a Malaysian city street. Ground violently shakes, he nearly falls, buildings shake and glass explodes. Concrete crashes behind him. He looks back briefly and accelerates. 00:03–00:06- COLLAPSE Front tracking shot racing backward. Building façade collapses behind him. He sharply changes direction, jumps over a metal barrier as concrete blocks destroy his original path. Dust fills the street. 00:06–00:09- CAR IMPACT He reaches an intersection. Vehicles slide uncontrollably toward him. He dives at the last second. Camera dives with him. A car passes overhead. He rolls, immediately gets up and keeps running. 00:09–00:12-ROAD CRACK A huge fissure tears across the road. He runs onto a damaged car, uses it as a springboard and leaps across the collapsing pavement. Lands hard and continues running. 00:12–00:15- FALLING GLASS Side tracking shot. High-rise windows explode above him. Glass rains down. He shields his face, ducks beneath a falling metal sign and keeps moving. 00:15–00:18-COLLAPSE A building front collapses ahead. He spots a narrow gap beneath the debris, drops low and slides underneath. Concrete crashes above him. He emerges dusty and immediately runs. 00:18–00:21-VEHICLE CHAOS A bus crashes into cars. A spinning vehicle comes toward him. He jumps onto its hood, springs off the roof and lands on the opposite sidewalk as sparks erupt behind him. 00:21–00:24- ELEVATED ROAD FAILURE An elevated roadway breaks apart ahead. He sprints to the edge and performs a full-body gap jump over the collapsing section. Brief dramatic slow motion, then real-time. Hard landing, roll, recover, run. 00:24–00:27- DEBRIS CHASE Heavy dust reduces visibility. Close shoulder-level pursuit. Falling streetlights and concrete force sudden direction changes. He stumbles against a wall, pushes himself forward and keeps running. A distant vehicle explosion sends a dust shockwave toward him. 00:27–00:30- FINAL COLLAPSE Camera moves ahead, facing him. Massive buildings collapse behind him as he runs toward camera. He jumps over a concrete barrier. A gigantic slab crashes behind him, creating a huge dust cloud. He emerges through the dust still running. CUT TO BLACK. One final footstep. SOUND: Deep earthquake rumble, structural cracking, glass explosions, crowd panic, tires, metal impacts, concrete crashes, heavy breathing, sparks, distant explosion and final low-frequency collapse. PERFORMANCE: He behaves like a real person fighting to survive not an action hero posing. Fear, instinct and adrenaline control every reaction. Realistic balance, impacts and recovery. CAMERA: Low tracking, front pursuit, side tracking, shoulder-mounted chase, whip-pans, ground-level slides and brief airborne follow shots. Camera always follows his momentum. No random drone shots, floating cameras or impossible rotations. FINAL FEEL: Relentless 30-second survival adrenaline — RUN → COLLAPSE → EVADE → JUMP → IMPACT → RECOVER → RUN → BIGGER COLLAPSE → RUN. Duration: 30 seconds | Aspect Ratio: 16:9
Details
0:20
ByteDanceSeedance 2.5
Image1 = woman holding gun from behind; Image2 = woman in foreground receiving drink. Strictly keep both women's faces, hairstyles, clothing, and identities consistent — do not swap them. Generate a 20-second, 16:9, live-action cinematic, one continuous take, no BGM crime-suspense deadpan comedy short film. Modern cold gray indoor beverage area, silver soda machine, cold white overhead lights. Camera fixed at a low angle slightly in front of and beside the soda machine, roughly waist-to-chest height, 35–50mm, medium close shot, shallow depth of field, essentially static. 0–4s Image2 character B cautiously approaches the soda machine, movements restrained and natural, only one brief wary glance around, neutral expression, no smiling. 4–7s She places a transparent plastic cup under the dispenser nozzle, only uses peripheral vision to check behind her once, then focuses on the cup. 7–12s Orange-yellow drink flows into the cup and gradually fills it, the machine fails to stop in time. The drink must realistically overflow past the cup rim → run down the cup wall → flow over her fingers → drip and splatter onto the metal tray. B only slightly furrows her brow, purses her lips, tightens her fingers — a sub-one-second "oh no" reaction, not exaggerated. 12–14s | Key reversal The camera does not search for danger. A black handgun slowly enters the existing frame from B's right rear, muzzle approaching the side-back of her head; character A should not appear fully yet at this moment. B reacts: her gaze first stops → extremely slight confirmation glance to the right rear → expression instantly goes quiet, body freezes. No wide eyes, no open mouth, no screaming, no jerking her head around. 14–17s B stays facing the soda machine, very slowly stops handling the cup and gradually raises both hands. Camera position unchanged, begins one natural cinematic rack focus: foreground B gradually falls out of focus, A at the right rear gradually sharpens. 17–20s Image1 character A becomes the sharp subject: black straight long hair, gold long earrings, black one-piece swimsuit, one hand steadily holding the gun, expression completely calm and neutral — no sneer, no menacing face. She looks at B and speaks in an ordinary conversational low voice, calm and earnest: "You know too much." B keeps her hands raised, does not turn around, does not respond — only an extremely subtle pause in her gaze. The shot ends immediately after the final word of the dialogue. Performance principles: B: restrained alertness → intense focus → brief hesitation → instant freeze. The greater the danger, the less she performs. A: always calm, steady, neutral — the threat comes from the act of holding the gun, not her expression. **Sound:** Only real production sound: footsteps, plastic cup, soda machine, liquid flowing and overflowing, splatter on metal tray, fabric rustling, and the final Chinese dialogue. No music, no horror sound effects, no cocking sound.
Details
0:30
ByteDanceSeedance 2.5
GIRL = MAIN CHARACTER: Young Korean woman in her early 20s, natural makeup, black wavy hair tied into a messy side ponytail, loose peach T-shirt, white pajama pants, black sneakers and simple necklace. Maintain exact identity and appearance. LOCATION: Small Korean apartment courtyard during a warm windy afternoon. Concrete walls, laundry lines, potted plants, bicycles and neighboring homes. CAMERA / LOOK: Early-2000s consumer DV camcorder. Strong handheld shake, autofocus hunting, exposure pumping, faded colors, soft contrast, motion blur and imperfect zooms. SEQUENCE: She hangs freshly washed clothes on a long outdoor clothesline. Everything looks calm until a sudden strong gust sends one shirt flying off the line. She runs after it while laughing. The camera operator follows shakily. The shirt lands on a parked bicycle. She grabs it, tries to hang it again, and another piece of laundry immediately blows away. She looks toward the camera in disbelief before chasing it down the lane. AUDIO: Wind, fabric flapping, birds, footsteps, distant scooters and neighborhood voices. No music. Candid domestic humor, realistic fabric movement, nostalgic DV imperfections and believable everyday chaos.
Details
0:30
ByteDanceSeedance 2.5
30-Second Cinematic Ultra-photorealistic Korean summer lifestyle video following a young woman as she unlocks her city bicycle, places her crossbody bag in the front basket, and naturally rides through a peaceful residential neighborhood. She slows down, turns realistically, parks beside the grocery store, locks her bicycle, and enters the supermarket. She casually selects a few groceries—apples, milk, and a snack—then pays at the checkout. After placing the groceries securely in the bicycle basket, she unlocks the bike, mounts naturally, and cycles home along the same street. Focus on realistic human movement, bicycle balance, pedal motion, braking, tire contact, momentum, wind, clothing and hair movement, grocery-bag weight, and authentic everyday sounds. Smooth cinematic camera movement, natural lighting, realistic depth of field, and continuous cause-and-effect action. No teleportation, instant acceleration, floating objects, distorted anatomy, changing clothes or bicycle, exaggerated wind, CGI appearance, or action-movie movement.
Details
0:30
ByteDanceSeedance 2.5
Multi-shot sequence with hard cuts between shots. Keep the same characters and visual style throughout. No music or score — only natural diegetic sound. SHOT 1 A dirty, exhausted 18-year-old girl with dark auburn-brown hair in a messy low ponytail, warm brown eyes, faded olive-green military jacket, backpack and gloves, aims a bolt-action rifle inside an abandoned ornate mansion. Sweat, dirt, a cheek scrape and tears cover her face. Her hands tremble. She closes her eyes as a tear rolls down, then hears something and lowers the rifle, visibly heartbroken. SHOT 2 Over her shoulder, an infected golden-coated Golden Retriever sits on the debris-covered parquet floor. Matted fur, cloudy white eyes, wounds and dark veins. He trembles while softly whimpering and growling. SHOT 3 She slowly crouches, places the rifle on the floor, wipes her tears and gives the dog a small broken smile. She quietly speaks to him, pulls a worn yellow-green tennis ball from her jacket and gently rolls it toward him. SHOT 4 Extreme low-angle close-up of the tennis ball rolling through dust across the cracked parquet floor. It stops directly in front of the dog's paws. Hold on the still ball. HARD CUT / TRANSFORMATION SHOT 5 Same low camera angle and ball position, but now on a clean sunny residential basketball court. A healthy golden-coated Golden Retriever suddenly grabs the ball. His eyes are clear and warm brown, fur clean, playful and energetic. He shakes and plays with the ball. SHOT 6 Wide shot of the basketball court. The woman throws the ball and the happy dog sprints after it, catches it, runs back and drops it at her feet, ready to play. SHOT 7 Close-up of the clean, unharmed woman laughing while affectionately ruffling the dog's neck and chest. The dog happily pants beside her. SHOT 8 Wide peaceful shot. The woman and dog sit together in a pristine dark-green vintage convertible beside the court. She shares food from a tin can with him. He gently takes it and licks her fingers. CONSISTENCY Same woman throughout: dark auburn-brown hair, loose low ponytail, warm brown eyes, olive-green military jacket, khaki cargo pants and brown boots. Same Golden Retriever: infected and injured in the mansion, healthy and playful after the transformation. Mansion: abandoned ornate interior, tall arched windows, torn dark-red curtains, peeling damask wallpaper, shattered gilt mirror, broken furniture, debris, ceiling hole and strong sunlight beams. Court: clean residential basketball court, white lines, iron fence, flowering shrubs, stone townhouses, warm late-afternoon sunlight and soft shadows. Dark-green vintage convertible with chrome trim. Visual style: ultra-realistic, photorealistic, natural human and animal movement, handheld camera, shallow depth of field, ARRI Alexa 35 look, warm cinematic grade, subtle grain. Audio: ONLY natural breathing, house ambience, dog sounds, footsteps, ball sounds, birds and wind. ABSOLUTELY NO MUSIC OR SCORE.
Details
0:10
ByteDanceSeedance 2.5
WHITE DRAGON ASCENSION | 10s | 16:9 Use image1 as SANZO. Keep his exact face, hair, skin tone, physique, tattoos and outfit. Ultra-realistic live-action Hollywood action. Practical stunts, real katana combat, realistic physics, natural smoke and cinematic lighting. No CGI, anime, cartoon or game look. LOCATION: Infinite white cyclorama with glossy reflective floor, soft diffused light, no horizon or props. OUTFIT: All-white tactical shinobi outfit with hood, half-mask, layered robes, armor, belt, tabi boots and flowing fabric. WEAPON: ONE white-handled katana on LEFT WAIST, fully sheathed. SANZO draws it once. Blade and handle leave the saya together. Saya becomes EMPTY and remains visible. No duplicate weapon or handle. Katana never returns. CREATURE: Giant aggressive celestial dragon made of dense swirling white smoke, with horns, whiskers, mane and bright white eyes. 0–2s: SANZO runs through the white void as the smoke dragon chases him. FPV and low tracking shots. Jaws snap behind him. He combat-slides beneath the jaws. Slow motion, eye close-up, 2-frame freeze. 2–3.5s: During the slide, SANZO draws the katana. Empty saya stays visible. He performs cartwheel → butterfly twist → forward somersault and lands low. 3.5–8.8s: SANZO unleashes KAGE RYUJIN GIRI — CELESTIAL EXECUTION. The dragon coils, dives, bites and tail-whips. SANZO counters with flash steps, flips, twists and aerial slashes. Each slash tears through the smoke; the dragon instantly reforms. Use fast camera moves, crash zooms, speed ramps and heavy impacts. 8.8–10s: Final dive. SANZO performs an aerial corkscrew, somersault and butterfly motion, then delivers a powerful upward diagonal slash. 2-frame freeze. Dragon breaks into white smoke ribbons and disappears. SANZO lands in a strong ninja stance, katana in RIGHT hand, blade down. EMPTY saya visible on LEFT WAIST. Fade white. NEGATIVE: CGI, anime, cartoon, game graphics, glow, beams, duplicate katana, duplicate handle, sheathing, clipping, floating weapon, blur, plastic skin, watermark, logo, subtitles, text.
Details
0:30
ByteDanceSeedance 2.5
Create a 30-second, 1080p ultra-realistic personal home-video showing an ordinary summer evening in the life of a young Japanese woman. No reference image. MAIN SUBJECT Young Japanese woman in her early 20s, naturally pretty, realistic skin texture, minimal makeup, soft and cheerful personality. Shoulder-length dark brown hair worn down with a simple clip pinning back one side, a few loose strands framing her face. Wearing an oversized light-yellow short-sleeve shirt, denim shorts, white canvas sneakers and a thin gold anklet. Maintain the same face, hairstyle, clothing, body proportions and overall appearance throughout the entire video. SETTING A quiet older residential neighborhood in Kyoto during a warm summer evening. Narrow paved streets, low wooden houses, hanging laundry, vending machines, telephone poles, tangled overhead wires, a small corner convenience store, a shrine gate visible in the distance, cicadas humming in the trees, and long tree shadows stretching across the road. Everything should feel lived-in, ordinary and peaceful. No tourist attractions, advertisements, recognizable brands or commercial activity. CAMERA / VISUAL AESTHETIC Raw personal footage casually recorded by a friend on an early-2000s consumer DV camcorder. Strong handheld shake, imperfect framing, autofocus hunting, exposure shifts, occasional motion blur, faded colors, soft digital detail, mild noise, accidental zooms and natural camera imperfections. No stabilization, drone footage, gimbal movement, dramatic lighting or polished commercial cinematography. — OUTSIDE THE HOUSE She steps outside holding a small paper fan. She slides the door shut, glances at her reflection in a window, and looks toward the camera with a shy laugh. A warm evening breeze rustles the nearby wind chimes as she starts walking down the street. — CONVENIENCE STORE STOP She reaches the small corner store and buys a popsicle. She steps outside, unwraps it, takes her first bite and scrunches her face happily at the taste. She notices her friend filming and waves the popsicle toward the camera with a grin before taking another bite. — WALKING HOME She continues down the quiet street while eating. A small dog tied outside a shop barks softly and wags its tail. She kneels down and gently scratches behind its ears. The dog's owner calls it back and she stands, giggling as she walks on. — SMALL ACCIDENT She walks past a low hanging laundry line when a stray sheet brushes against her head. She stops, startled, then realizes what happened and bursts out laughing. She pats down her hair while her friend zooms clumsily toward her face. She looks directly into the camera and says: "Mou, yamete!" ("Come on, stop it!") — QUIET MOMENT She reaches a small stone step near a shrine gate and sits down for a moment. She finishes the popsicle while watching fireflies begin to flicker faintly near the bushes. A gentle breeze moves her hair. She tucks a loose strand behind her ear and looks around peacefully. — FINAL MOMENT She stands up and continues walking home. After a few steps, she turns back toward the camera with a soft playful smile and says: "Hayaku kite." ("Come on, hurry up.") She keeps walking down the dim lane as the streetlights flicker on, while the camera follows behind. She disappears around the corner and the recording abruptly cuts to black. AUDIO Natural location sound only: cicadas, wind chimes, footsteps, distant bicycle bells, a vending machine hum, sliding doors, rustling laundry, faint conversations and subtle camera-handling noise. No music. No narration. No artificial sound effects. Only the two spoken lines specified above. FINAL FEEL The entire video should feel like a forgotten summer recording from the early 2000s. Warm, gentle, casual, slightly playful, intimate and nostalgic. No dramatic events, no exaggerated acting, no model poses and no polished cinematic shots. The realism should come from ordinary surroundings, natural expressions, small accidents, realistic environmental movement and imperfect camcorder operation.
Details
0:15
ByteDanceSeedance 2.5
Concept: "Golden Hour in the Old Town" Create a hyper-realistic observational video filmed on a modern mirrorless camera with a handheld documentary style. The camera is never visible, and the woman is not recording herself. The footage is captured naturally by someone walking nearby, creating the feeling of an authentic lifestyle documentary rather than a vlog or commercial. Use realistic handheld camera movement with subtle footsteps, gentle breathing, occasional autofocus adjustments, natural exposure adaptation, slight framing imperfections, and authentic environmental sound. Avoid cinematic camera moves, slow motion, transitions, or dramatic editing. Everything should feel like genuine candid footage. Visual Style Modern documentary-style footage with natural colours, balanced dynamic range, soft golden-hour sunlight, realistic skin texture, and accurate exposure. No beauty filters, excessive sharpening, cinematic colour grading, lens flares, or artificial effects. Character A beautiful young European woman in her twenties with long wavy chestnut hair, fair skin, and elegant natural makeup. She wears a fitted white crop top, high-waisted beige linen shorts, white sneakers, delicate gold jewellery, sunglasses resting on her head, and carries a woven shoulder bag. Her appearance remains consistent throughout. Location A charming historic European old town with narrow cobblestone streets, cafés spilling onto the sidewalks, flower-covered balconies, cyclists, outdoor markets, and warm sunlight reflecting from centuries-old stone buildings. The recording begins as she exits a small artisan bakery carrying a paper bag of fresh pastries and an iced coffee. She pauses for a moment to enjoy the warm afternoon sun before continuing down the cobblestone street. She stops outside a flower stall, smiles while smelling a bouquet of fresh lavender, and exchanges a friendly smile with the florist before continuing. As she walks through a lively square, a violinist performs nearby. She slows down, listens for a few seconds, gently applauds with a genuine smile, and drops a coin into the musician's case before walking away. A light breeze moves her hair as pigeons scatter across the square. She glances back toward the unseen camera operator with a warm smile and a small wave before continuing toward a narrow sunlit alley lined with cafés and climbing ivy. The camera lingers for a moment as she gradually disappears into the bustling old town, ending the recording naturally. Generation Requirements Produce one continuous 15-second handheld recording with consistent facial identity, realistic anatomy, authentic walking mechanics, believable interactions, natural eye movement, and physically accurate lighting. Include realistic ambient sounds such as footsteps on cobblestones, distant conversations, birds, café activity, bicycles, and soft street music. Avoid AI artifacts, duplicated pedestrians, warped anatomy, floating objects, perfect symmetry, cinematic effects, or unnatural motion. The final result should be indistinguishable from genuine documentary footage casually filmed during an afternoon in a European city.
Details
Browse all Seedance 2.5 promptsExplore AI Video Prompts