cinematic martial arts confrontation in broad daylight, a blind shaolin monk wearing a dark, stylized combat outfit inspired by legendary fighters stands calm and centered, eyes closed, surrounded by multiple hostile creatures emerging from a traditional Japanese landscape
Ultra cinematic choreography coverage, mix of slow dolly-ins + orbit moves + whip pans, transitions masked by body motion and impacts, alternating real-time and slow motion, continuous fluid sequence
(0-2s) wide establishing shot, monk standing still in center, wind moving fabric, creatures circling, tension builds
(2-4s) slow push-in close-up on monk’s face, eyes closed, subtle head tilt sensing movement
(4-6s) sudden attack from first creature, monk reacts instantly, precise sidestep + redirection, fluid motion
(6-8s) chained combat sequence, monk engages multiple opponents, spinning strikes, controlled movements, each impact sending creatures flying backward with stylized motion
(8-10s) slow motion highlight: mid-air dodge + counter sequence, cloth movement and body rotation emphasized, creatures suspended briefly before being thrown away
(10-12s) final burst of speed, monk flows through remaining opponents in one continuous movement, camera orbiting rapidly, enemies collapsing or being thrown aside
Traditional Japanese environment, open landscape with temples, wooden structures, distant mountains, clear daylight, subtle wind movement, dust and debris reacting to motion
Ultra realistic, high-end martial arts film choreography, precise body mechanics, cinematic slow motion, strong contrast lighting, volumetric atmosphere, fluid transitions, intense but controlled physical interaction, no distortion, no stretching
cinematic martial arts confrontation in broad daylight, a blind shaolin monk wearing a dark, stylized combat outfit inspired by legendary fighters stands calm and centered, eyes closed, surrounded by multiple hostile creatures emerging from a traditional Japanese landscape
Ultra cinematic choreography coverage, mix of slow dolly-ins + orbit moves + whip pans, transitions masked by body motion and impacts, alternating real-time and slow motion, continuous fluid sequence
(0-2s) wide establishing shot, monk standing still in center, wind moving fabric, creatures circling, tension builds
(2-4s) slow push-in close-up on monk’s face, eyes closed, subtle head tilt sensing movement
(4-6s) sudden attack from first creature, monk reacts instantly, precise sidestep + redirection, fluid motion
(6-8s) chained combat sequence, monk engages multiple opponents, spinning strikes, controlled movements, each impact sending creatures flying backward with stylized motion
(8-10s) slow motion highlight: mid-air dodge + counter sequence, cloth movement and body rotation emphasized, creatures suspended briefly before being thrown away
(10-12s) final burst of speed, monk flows through remaining opponents in one continuous movement, camera orbiting rapidly, enemies collapsing or being thrown aside
Traditional Japanese environment, open landscape with temples, wooden structures, distant mountains, clear daylight, subtle wind movement, dust and debris reacting to motion
Ultra realistic, high-end martial arts film choreography, precise body mechanics, cinematic slow motion, strong contrast lighting, volumetric atmosphere, fluid transitions, intense but controlled physical interaction, no distortion, no stretching
SUBJECTS / PEOPLEMale protagonist, in his 30s, cold and tough, short hair, sharp facial features, wearing a well-fitted black suit, black shirt, and black tie. The outfit remains consistent throughout the entire video. His movements are steady, precise, and highly professional. He has extremely strong body control, with short and steady steps. In close combat, his movements are like a highly trained tactical fighter: turning, gun-suppression, elbow strikes, deflecting force, and repositioning—all executed smoothly in one breath. The enemies are a group of adult male thugs, wearing a mix of dark street clothes or suits, with clear primary and secondary roles. They are responsible for rushing out from all sides and approaching in turns, without stealing the visual focus from the protagonist. The male lead is always the center of the frame, and all actions revolve around him. His performance exudes a calm, ruthless aura with no wasted movements.
OBJECTSA black tactical pistol. It is always the same gun—no morphing, no disappearing, and no second gun ever appears. Enemies can approach with melee weapons or bare-handed, but there should be no exaggerated weapon stacking and no heavy firearms. Knocked-down enemies can gradually appear on the ground, but the change in numbers must be logical; all fallen individuals come from the same group that rushed out earlier. The male lead's gun is always clearly visible. During close combat, the gun remains in his hand or pressed near his body line; it should never suddenly disappear. No extra weapons, no duplicate hero, no duplicated enemies, same pistol throughout the full shot sequence.
ENVIRONMENTAn indoor hall at night or an underground club corridor. The space is wide but oppressive, with a black and gold color palette. The floor is slightly reflective. The background features pillars, dark walls, localized warm lights, and cold rim lighting. The foreground of the scene leaves room for the protagonist's actions, the midground shows constantly approaching enemies, and the background reveals more silhouettes rushing out from the shadows. Lighting shines down from above and from the side/rear, creating strong contrasting silhouettes, ensuring the male lead's suit, pistol, and body turns are clearly visible. Environmental elements remain consistent throughout the video. Do not cut to unrelated spaces, and do not suddenly change to a street or rooftop.
MOODHigh-pressure, cold-blooded, explosive. It is not an out-of-control brawl, but a precise, slaughterhouse-style suppression. The pacing gets increasingly fierce, but remains legible at all times. Emotionally, it must deliver the oppressive thrill of "facing a multi-person siege alone, yet completely controlling the situation."
CAMERA LANGUAGEThird-person cinematic camera, mainly using medium-close shots and tight wide shots, interspersed with a few close-ups to emphasize the moment of takedowns. The camera overall uses a stabilizer follow-shot + a slight handheld impact feel. Cut as little as possible, prioritizing continuous panning, pushing forward, and orbital tracking shots to reinforce the illusion of a one-take. The camera always revolves around the male lead. It never loses track of the main gun, never loses the protagonist's body orientation, and never lets enemies completely block the frame. Every takedown action must be clearly seen: who rushes up, how the male lead makes contact, how he switches fire, how the enemy falls, and where the camera finally rests.
CONSISTENCY RULESThe male lead's clothing, hairstyle, and body shape remain unchanged throughout. There is always only the same male lead; no stunt double faces, no duplicate protagonists. There is always only the same black pistol. The enemies come from the same group that rushed out; the decreasing number is logical, and they do not suddenly respawn. Every takedown must clearly show the contact, counterattack, and the resulting fall. No chaotic blind firing, no floating movements, no supernatural abilities. The male lead consistently maintains a professional, calm, and precise rhythm. The ending must be a final display shot of the male lead standing steadily, surrounded by fallen enemies, with his gun muzzle pointing down or resting at his side.
TIMELINE0:00–0:02A tight medium shot starts from behind the male lead's side, and the camera slowly pushes forward. The male lead stands in the center of the hall, slightly turned, shoulder line tense, holding the same black pistol, his gaze locked on the shadows ahead. In the background, several enemies suddenly rush out from both sides and directly in front, their footsteps syncing with heavy drumbeats. The camera stops at a position where both the male lead's half-profile and the incoming crowd ahead are clearly visible simultaneously; the crisis has formed.
0:02–0:05The camera quickly pushes in to a medium-close shot. The first enemy lunges forward into close-combat distance. The male lead side-steps to dodge the trajectory, raises his hand to press down on the enemy's shoulder/neck, uses the pistol to maintain close-line control, and executes a clean, sharp close-range counterattack. The first person immediately loses balance and smashes into the ground. At almost the same rhythm, the male lead uses the momentum to turn around. A second enemy rushes in from the side; he uses an elbow strike and the gun body to create space, then follows up with an extremely short finishing move, flipping the second person right at his feet. The camera follows the male lead's turn with a slight pan, finally resting on a state where two enemies are already down, and a third is about to enter the frame.
0:05–0:08The camera slightly orbits to the male lead's front, maintaining a medium shot so his full body actions are legible. The third and fourth enemies approach one after the other. The male lead first steps forward, uses a close-quarters block to interrupt the former's action, then immediately switches fire to handle the latter. The movements are like a compressed spring instantly releasing—clean, fast, with no extra swaying. The former staggers back after being interrupted, while the latter is directly taken down. The male lead immediately turns back to deliver the final suppression on the former, completing a continuous double kill. The camera stabilizes as the male lead returns to a neutral posture, clearly showing multiple people already fallen around him.
0:08–0:11The camera slightly pulls back to a wider medium shot, showing the spatial relationship. The remaining several enemies close in simultaneously from both the left and right flanks. The male lead does not retreat; instead, he cuts into the most concentrated side, using extremely short footwork to continuously shift positions, pulling the enemies one by one into his close-combat rhythm. He first controls the closest person to use as a brief shield, then quickly spins out from the edge of that shield to take down two people in succession. The last one tries to lunge at him from behind but is precisely intercepted by the male lead turning around and slammed into the ground. The camera makes a clean arc movement as he cuts in and turns back, ending at the moment the male lead stands back in the center, with the last still-moving enemy staggering backward.
0:11–0:13A quick push-in to a close shot, focusing on the final confrontation. The last enemy roars and charges head-on. The male lead's eyes are calm, his body sinks slightly, and he waits for the opponent to enter range before instantly striking, taking him down with the sharpest set of close-quarters gun-fu. This action must be fully visible: contact, control, counterattack, and falling down, all in one smooth motion. The camera shakes slightly when the enemy hits the ground heavily, then quickly stabilizes, clearly showing the male lead still standing.
0:13–0:15The camera slowly pulls back to a heroic medium-wide shot. The male lead stands in the center of the fallen crowd, breathing steadily, his suit still immaculate, holding the same black pistol, the muzzle naturally pointing down or resting at his side. All surrounding enemies have fallen and cannot get up. The background lights and dust float slightly as the drumbeat hits its final note. The final shot freezes on the moment the male lead coldly raises his eyes to look at the camera, forming a strong, clean, and memorable final frame.
{
"model": "seedance-2.0",
"task": "generation",
"type": "directed_video",
"metadata": {
"scene": "A brawl in a super-wide subway car - A woman in a tracksuit fights two casually dressed thugs in a cramped subway car. Using her speed and agility, she confronts the thugs, stands on their fallen opponents, and emerges victorious.",
"theme": "The energy of David and Goliath - extreme size difference, a woman's innocence versus brutal violence, intense impact, and maintaining composure until the very end despite being out of breath."
},
"inference": {
"use_reference": "@rio_ジャージ_1.png ",
"description": "A Japanese woman with silver-platinum hair tied in a high ponytail. She is wearing a green tracksuit, a black tutu, and black boots. She is petite and has a stern expression. Her opponents are two muscular, casually dressed delinquent men. One has blond hair and dark skin, while the other has scars, lighter skin, and is wearing khaki cargo pants. Both are taller and heavier than her, and it is evident that they are covered in blood and bruises from the fight."
},
"sound": {
"type": "diegetic only",
"levels": {
"heavy bare-knuckle impacts - wet thuds of fists on exposed skin": "constant",
"girl's sharp exhale grunts on each strike - short, controlled": "constant",
"large bodies crashing into seats and handrails - metallic clang and plastic crack": "constant",
"swift fabric whip sound on spinning kicks": "constant",
"subway car rumble and mechanical rattle - constant",
"overhead handles swinging and clattering after impacts",
"heavy breathing and groaning from wounded fighters on the floor"
}
},
"camera": {
"framing": "dynamic low-angle tracking, ultrawide 16:9 framing",
"movement": "fluid steadicam-style tracking down the center aisle of the car, whip pans on direction changes, drops lower for ground-level strikes, pulls to wide symmetrical composition for 2-on-1 flanking moments"
},
"environment": {
"setting": "interior of a japanese-style subway car - blue molded seats transitioning to burgundy/red seats in adjacent car sections, chrome poles, overhead grab handles, fluorescent ceiling strip, ad placards, sliding doors",
"lighting": "Cold fluorescent lights shone overhead, emitting a slightly greenish tint, and the strong skylight cast deep shadows on the thugs' eyes and chests, while warm orange reflected light from the tunnel flickered outside the windows.",
"atmosphere": "The train car was empty—civilians were huddled in the corners, some slumped unconscious in their seats, and the floor was slightly soiled."
},
"action": {
"0:00-0:06": "The woman is already engaged in a fierce battle with the first large opponent, dodging his heavy attacks with quick movements and landing a combination attack to his torso.",
"0:06-0:15": "The 2-on-1 fight continues inside the car. Combination attacks of aerial kicks, knee strikes, and elbow strikes using the handle from above as leverage. The woman continues to fight despite being attacked, and the men suffer visible damage: cuts, bruises, staggering, etc." },
"style_controls": {
"duration": "15 seconds",
"vfx": "none",
"feeling": "brutal grounded action, korean action cinema aesthetic - hard impacts with physical consequences, no superpowers",
"quality_guardrails": "No flickering, no distortion, consistent costume details throughout all frames (jersey design, hairstyle, tutu skirt, boots), a stable number of opponents - exactly two delinquent men in casual fashion, no extra limbs during quick kicks, consistent sound synchronization - injuries accumulate but don't disappear, no passing through seats or pillars, hair remains in ponytail."
}
}
15-second Original Desert Martial Arts Short Film: A black cat warrior in light armor stands alone in a desert where yellow sand is flying all over the sky, facing the pursuers. The shots combine slow motion and fast editing; under backlight, the yellow sand rolls like ink mist. The character's movements are elegant yet ferocious, with tattered but flowing robes. Holding a short weapon, he shuttles and counterattacks at high speed. The overall tone is cold, lonely and oppressive, with high-end colors and obvious shallow depth of field, just like a high-quality oriental martial arts movie.
Seedance 2.0
