Prompts
Featured
Models
Submit
Blog

AhaPrompt

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

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 Policy Terms of Service

HomeAI Prompt LibraryAI Video PromptsUnderwater 3D Kids Animation

Underwater 3D Kids Animation

Seedance 2.5ByteDancevideo

Create a premium colorful underwater 3D kids animation. Main character: Cute toddler girl wearing a magical diving suit. 0–3 sec: A glowing bubble transports her underwater. 3–6 sec: Friendly fish swims past. Display and said: “FISH 🐟” 6–9 sec: Cute turtle approaches. Display and said: “TURTLE 🐢” 9–12 sec: Smiling dolphin jumps through the water. Display and said: “DOLPHIN 🐬” 12–15 sec: All swim together around colorful coral. Display and said: “Ocean Friends! 🌊”

Details

Type: AI Video Prompts
Aspect Ratio: 9:16
Duration: 0:10
Updated: Oct 7, 2026

Category

Animation Cartoon 3D Render

How to use this prompt

1Copy the full prompt with the copy button above.
2Open Seedance 2.5 by ByteDance, or your preferred AI video generator.
3Paste the prompt and keep the camera and motion terms — they steer the cinematography.
4Match the example's settings (9:16 · 0:10 · Animation · Cartoon · 3D Render), 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: 9:16 · 0:10 · Animation · Cartoon · 3D Render. 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

Create a clean, premium 8-second fashion infographic animation based on the outfit-board layout. The entire composition remains visible on a white minimalist background.The female model on the right begins a smooth 360° rotation on her vertical axis, like a fashion showcase turntable. Her movement is elegant and continuous, maintaining natural posture and realistic cloth physics. Hair, skirt straps, and accessories react subtly to motion. At the exact same time, every item displayed inside the numbered boxes rotates synchronously:Necklace rotates slowly in 3D, revealing chain depth and pendant details. Hair bow spins gracefully around its center.Handhel d gaming console rotates on its vertical axis like a product commercial.Cargo mini skirt rotates 360° to showcase front, side, and back views.Platform boots rotate together like luxury product display renders. Black off-shoulder top rotates smoothly to reveal full garment shape.All items complete their rotations in perfect sync with the model's rotation speed. The boxes, numbers, typography, and layout remain fixed in position while only the products rotate inside their frames. Camera remains mostly static with a subtle cinematic push-in. Soft studio lighting, realistic shadows, clean commercial fashion-ad aesthetic, ultra-sharp details, premium motion graphics, smooth easing, luxury catalog presentation, 4K, 60fps.

Strictly reference the female kayaking athlete, deep navy-blue whitewater kayak, black paddle shaft with bright orange blades, life jacket, short white hair, canyon waterfall environment, and the overall high-end painterly animation visual style from the input video. Maintain extremely high consistency of the character’s face, hairstyle, clothing, body proportions, kayak structure, hull color, double-bladed paddle, and canyon environment throughout the entire 30 seconds. Do not redesign the character. The core story follows an experienced female kayaker who independently challenges a dangerous waterfall canyon rapid. She enters a narrow canyon from relatively calm upstream water, rapidly paddles through continuous rapids, avoids large rocks, breaks through wave walls, then descends a small drop. During the final section, she crosses the most violent whitewater before finally entering a relatively calm deep-blue pool beneath the waterfall. The entire sequence emphasizes realistic kayaking technique, the power of the water, speed, danger, and the athlete’s calm control. No dialogue, no text, no logos, no watermark. 0–4 seconds | Establishing the waterfall canyon An ultra-wide establishing shot reveals a massive rocky canyon. Towering gray-blue vertical cliffs rise on both sides. Multiple waterfalls cascade down from a great height in the distance, while large amounts of white mist float through the air. A deep-blue river flows through the bottom of the canyon, its surface covered with white foam and rapids of different sizes. Sunlight enters from above the canyon, creating bright highlights on the wet rocks and water. The camera slowly dives downward from a high position, flying along the direction of the river and into the canyon. It eventually reveals the female athlete driving a deep navy-blue single-person whitewater kayak into the frame from a distance. She has short white hair, blue eyes, a dark sleeveless athletic top, and a beige/tan life jacket. She holds a black double-bladed paddle with striking bright-orange blades. 4–8 seconds | Entering the first rapid The camera drops close to water level and moves backward in front of the athlete while tracking her. She quickly studies the water ahead. Her body leans slightly forward as she begins continuous alternating strokes. The right paddle blade enters the water. Her torso rotates while her arms pull the water backward, rapidly correcting the kayak toward the left. Immediately afterward, the left blade enters the water, bringing the kayak’s bow back toward the center of the rapid. Every time an orange paddle blade enters the water, it creates localized splashes, bubbles, and realistic resistance. The kayak visibly rises and falls with the waves while maintaining forward momentum. The camera briefly switches to a low side angle as the bow crashes into the first white wave wall, sending water exploding naturally to both sides and washing across the hull. 8–12 seconds | High-speed passage through the narrow rocks The river becomes progressively narrower, with enormous dark rocks appearing on both sides. The athlete sees a large rock directly ahead and immediately performs a powerful right-side stroke, forcing the kayak to cut left into the safer water channel. The camera switches to a high overhead view, clearly showing the kayak’s route through the complicated water flow. The main current moves rapidly downstream. Behind the rocks, realistic whirlpools and recirculating eddies form. The kayak rapidly travels along the edge of the white foam. The camera then moves close to the bow. The deep navy-blue hull races toward the camera while the orange paddle blades alternate from side to side. The hull occasionally tilts slightly due to turbulent water. The athlete immediately uses her waist and core strength to restore balance.

Create a high-quality cinematic 3D animated short film set in a vast sunny desert. A cute, chubby baby hedgehog-like desert animal with soft beige fur, tiny brown spots, huge expressive eyes, a small black nose, round cheeks, and an adorable innocent face sits beside an old dead tree near a desert highway. The environment is warm and colorful, with golden sand dunes, distant rocky mountains, scattered desert plants, cacti, and a clear bright blue sky. The tiny animal notices a large green cactus fruit/pad nearby and happily approaches it. Show close-up shots of its curious expression as it grabs and starts eating the cactus. The cactus has small thorns, while the animal struggles adorably but continues chewing with excitement. In the background, a tall camel slowly approaches the cactus. Cut to a cinematic close-up of the camel noticing the little animal and the cactus. The camel leans forward and takes a huge bite of the cactus, chewing it with exaggerated funny facial expressions. The baby animal looks shocked and confused as the camel eats its food. The camel suddenly spits some green cactus pulp back toward the little animal. Green pieces land all over the animal's face and fur. Show an extreme close-up of the baby animal covered in green cactus pieces, looking completely surprised. Its expression gradually changes from shock to happiness, with its tongue sticking out playfully and a cute smile. The baby animal then looks up at the camel. The camel reacts with a funny, smug expression. The little animal tries to reach toward the camel while the camel starts walking away. In the final sequence, the camel walks across the desert road while the tiny animal follows behind, stumbling and rolling adorably through the sand. End with a wide cinematic shot showing the camel and the tiny animal becoming smaller as they move down the long desert highway surrounded by golden dunes. Style: premium Pixar-like 3D animation, adorable stylized characters, highly expressive faces, cinematic composition, realistic desert environment, soft global illumination, detailed fur, natural shadows, vibrant colors, smooth character animation, comedic timing, dramatic close-ups, shallow depth of field, smooth camera movements, polished feature-film quality, 4K, vertical 9:16.

Create a 15-second vertical 9:16 cinematic 3D animated comedy scene in a warm, cozy bedroom. A cute chubby baby with blond hair, big blue eyes, and an orange shirt is playing beside his sleeping father, a young man with brown hair wearing a light-blue shirt. The baby first gently pokes and touches his father’s face while the father sleeps peacefully. The baby becomes mischievous and playful, climbs onto the bed, grabs a basket full of colorful toy blocks and teddy bears, and playfully bothers his sleeping father. The baby then leans down and playfully bites/nuzzles the father’s face, causing the father to suddenly wake up with a shocked expression and wide-open eyes. The father picks up the baby and looks at him seriously for a moment, creating a funny father-and-baby confrontation. Finally, both suddenly burst into laughter, with the baby laughing joyfully in his father’s arms. Style: high-quality Pixar-style 3D animation, adorable exaggerated facial expressions, realistic skin and hair details, soft cinematic lighting, warm golden sunlight through the window, cozy bedroom environment, detailed toys, smooth character animation, expressive body language, humorous timing. Camera: close-up facial shots, medium shots, gentle camera movement, quick comedic cuts, shallow depth of field, cinematic composition, smooth motion, consistent character appearance throughout.

1. A cute chubby toddler boy, around 2–3 years old, with fair skin, huge expressive green eyes, short golden-blond hair with a small curled tuft on top, rosy cheeks, and a mischievous smile. He wears a bright yellow one-piece pajama covered with tiny orange duck patterns and a blue zipper. 2. His father, a friendly middle-aged man with neatly styled dark-brown hair, thick eyebrows, round black glasses, a large brown mustache, and a warm but slightly strict expression. He wears a light-blue shirt with a darker blue sweater vest and brown pants. Environment: A bright, playful kindergarten classroom filled with pastel colors. Rainbow decorations, colorful clouds, stars, alphabet blocks, children's tables and chairs, toys, teddy bears, large windows, and soft warm daylight. The entire environment should feel cheerful, whimsical, safe, and child-friendly. Animation and story: Start with the toddler standing confidently in the classroom with a mischievous expression while his father stands behind him. The father looks suspicious and points toward the child as if questioning him. The toddler looks up at his father with an innocent expression and answers confidently. The father becomes more suspicious and asks whether the child has been eating sugar. The toddler tries to deny it while smiling mischievously. Show a close-up of the toddler laughing with his eyes closed, revealing tiny teeth and a playful expression. Cut back to the father and toddler together. The father raises one finger and questions him about telling lies. Then the father leans closer toward the toddler and gestures toward the child's mouth, telling him to open his mouth. End with an extreme close-up of the toddler laughing uncontrollably with eyes closed and a huge innocent smile while colorful confetti/star shapes appear around him. Visual style: High-end 3D animated children's movie quality, Pixar-inspired level of polish without copying any specific copyrighted character, soft rounded character design, oversized expressive eyes, adorable facial expressions, smooth skin shading, detailed clothing, vibrant pastel colors, soft global illumination, cinematic depth of field, volumetric lighting, glossy yet natural materials, beautiful subsurface scattering, clean character animation. Camera: Use dynamic cinematic camera movements with medium shots, close-ups, and expressive reaction shots. Smooth dolly-in movements, gentle camera transitions, shallow depth of field during close-ups, and stable framing. Keep the toddler's face highly expressive and clearly visible. Facial animation: Strong lip-sync, natural blinking, eyebrow movement, eye direction changes, cheek movement, smiling, laughing, surprised expressions, and exaggerated children's cartoon emotions. Motion: Smooth professional 3D character animation, natural hand gestures, believable body movement, subtle head movements, expressive pointing and reactions. Avoid stiff animation or unnatural movements. Lighting: Warm soft daylight coming through classroom windows, colorful ambient bounce light, soft shadows, subtle rim lighting, bright cheerful atmosphere. Audio: Cute child voice for the toddler, warm male father voice, playful children's background music, subtle classroom ambience, expressive comedic timing, clear dialogue and accurate lip synchronization. Dialogue timing: Father: “Johnny, Johnny!” Toddler: “Yes, Papa?” Father: “Eating sugar?” Toddler: “No, Papa!” Father: “Telling lies?” Toddler: “No, Papa!” Father: “Open your mouth!” Toddler: [laughs] Overall mood: Funny, adorable, innocent, playful, highly engaging children's animation with exaggerated facial expressions and comedic timing. Output: Ultra-detailed 3D animation, 9:16 vertical format, 4K quality, smooth motion, cinematic composition, consistent character identity throughout the entire video, professional children's animation, polished rendering.

The faces, bodies, forest, lightning, and destruction effects of the two giants consistently use the same clearly computer-generated AAA fantasy game cinematic aesthetic throughout, not live-action photography, live-action makeup, cosplay, ordinary gameplay, 2D animation, or cel shading; close-ups maintain digitally sculpted faces, bundled hair strands, and polished motion. Environment: a vast broadleaf forest under an overcast daytime sky, deep green canopy, gray-green storm clouds, restrained warm golden sunlight through a cloud gap at the rear right, distant blue-gray mountains, ground of wet brown soil, bare rock, thick tree roots, and broken branches. In the center of the forest is a wide east-west forest clearing, with dense woods connecting at both ends; all action takes place in this same clearing and the tree belts at its two ends, without cutting to a city, city walls, or an arena. The two giants are basically equal in height, the mature canopy only reaches around their waist and hips, tree trunks look small beside their legs, and a single tree crown is clearly smaller than a giant's chest; B has broader, thicker chest and shoulders, A is more slender. Wide angles continuously preserve trees, ground soil beneath their feet, and distant mountains to provide scale, and close-ups do not turn them into ordinary human size. Throughout the entire piece there are exactly two giants with adult male proportions, with no human soldiers, weapons, or additional giants. Giant A is fixed as the black-haired giant: shoulder-length messy black hair divided into thick bundles, a gaunt and fierce long face, deep eye sockets, prominent cheekbones, and a broad tooth-baring jaw, with both eyes having clear emerald green light, controlled in intensity, with eye direction visible. His build is lean but powerful, with broad shoulders, clearly defined chest and abdominal muscles, and thick arms; his whole body has complete biological skin from light gray-brown to pale flesh tone, with deep shadows between muscles, and continuous material between face and body. No clothing, armor, helmet, or weapons; the waist and hips are naturally obscured by foreground canopy, rear-side angles, and motion dust, without showing private parts, while the limbs and contact with the ground must still be clear. Both hands and bare feet are anatomically correct. Giant B is fixed as the white-haired armored giant: extremely broad shoulders, thick chest, thick limbs, a solid waist and hips, short dense silver-white hair with thick bundled strands, deep eye sockets, a fierce gaze, and a broad jaw, with eyes that do not emit green light. Rough bone-colored to dark golden-brown segmented biological hard armor covers his entire body, layered horizontally across the chest, with shoulder, arm, and thigh plates arranged along muscle direction, and irregular hard armor islands distributed across the back; between the armor plates, dark red muscle bundles and tendons are continuously exposed, which is his inherent complete biological appearance, not bleeding wounds. His face also has the same bone-colored hard structures, the lower jaw can open and close naturally for roaring, teeth are fixed; both hands and bare feet are complete, with narrow ring-like hard edges at the wrists and ankles. No coat, cape, helmet, shoes, or weapons. A's long black hair, pale complete skin, green eyes and B's short white hair, bone-brown biological armor, and dark red muscle bundles must remain distinct throughout, and must not mix, exchange, duplicate, or switch persons between cuts. Spatial and punching relationships are fixed: A starts from the woodland on the left side of the frame and charges right, B starts from the woodland on the right side and charges left; the first time, A's right fist knocks B back into the right-side tree belt, after which B counterattacks and knocks A back into the left-side tree belt. The camera mainly stays on the same side of this combat axis, using foreground tree trunks and terrain to maintain orientation. A uses only his right fist to complete the first horizontal strike and the subsequent missed downward smash; B hits A's ribs and abdomen with a left fist and immediately follows with a right uppercut. Each attack has a clear attacker, target, contact point, and force direction, and the two must not imitate the same action at the same time. 0.00–3.00 seconds: in the first frame, golden lightning strikes the left and right ends of the forest clearing respectively, the two strike points are about four giant-heights apart, leaving a completely visible clearing in the middle. Use one continuous low ultra-wide shot, from the lightning strikes until the two giants are revealed, always framing both strike points at the same time, with the camera only tilting upward, not cutting to another nearby patch of forest. A steam column beside the bare rock at the left end surrounds A, and a steam column beside the forked large tree at the right end surrounds B; A rises from a crouched posture at the original left strike point, B stands straight from a half-squat at the original right strike point, and heads, shoulders, chests, and arms emerge above their respective canopies. The feet of both giants always remain on the same ground at the bottom of their respective light columns, and after standing they are still four body-heights apart, with the bare rock, forked tree, and central clearing not moving. They are already giants to begin with, revealed by rising from behind the trees, without inserting small figures inflating. Thick bright golden main arcs branch along each person's shoulders, back, and chest, while thinner arcs rapidly travel along upper arms, forearms, and thigh surfaces and flash continuously several times; the electric light illuminates A's black hair, pale muscles, and green eyes, and illuminates B's silver-white hair, bone armor, and dark red armor seams, with local steam lit through by the electric glow. Bodies and faces remain clearly readable, canopies are bent by air pressure, leaves and steam curl outward from their respective strike points; before 3 seconds both are standing firmly, the main lightning between cloud and ground weakens, and the close-body arcs continue into the subsequent roar. 3.00–5.00 seconds: two quick connected low-angle chest-up shots, first A then B, both still remaining at their respective distant strike points. A faces the distant B to the right, long black hair thrown back, mouth open in a deep rough enormous roar, chest and neck tightening with exhalation; immediately cut to B facing the distant A to the left, opening his bone-colored lower jaw, answering with a heavier roar, fists clenched, shoulder armor tightening. Both close-ups preserve the canopy and steam background of their respective strike points, not letting the other giant suddenly stand beside their shoulder. Golden branching arcs continue flashing along chest, shoulders, and arms, then gradually extinguish before 5 seconds, while steam still drifts backward from their bodies. Both lean toward each other during the tail of the roar, eyes fixed on the opponent, not stopping to pose after roaring. 5.00–8.00 seconds: cut back to a side low-angle two-person wide shot that can simultaneously see both original strike points and the central clearing. A charges at full speed from the left end to the right, B from the right end to the left; both faces, chests, knees, and toes always face their respective direction of advance, upper bodies leaning toward the opponent, arms swinging powerfully with each stride. Each supporting foot briefly presses firmly into the ground, pushing soil backward behind them, driving the pelvis and entire torso forward; the other foot steps to a new point closer to the opponent, and only when the rear foot leaves the ground does mud flip up behind them. The camera's lateral position remains stable, only slightly pushing in, with ground bare rock, tree roots, and tree trunks at both ends serving as fixed references, without high-speed lateral foreground movement obscuring the feet. Clearly see both leaving their original steam columns, each continuously crossing new ground toward the center, their starting points growing farther behind them, and the gap between their chests continuously shrinking from four body-heights to punching distance. Preserve this continuous process of advancing until meeting, without replacing actual displacement with foot close-ups; no walking backward, no sliding with both feet, no running in place on a loop. Only when they approach the center does the camera cut closer, preparing the composition for the first contact. 8.00–12.00 seconds: during the approach, A first plants his front foot firmly on the ground, rotating hips and shoulders to drive his right arm into a powerful horizontal punch; B urgently plants his front foot to prepare to stop the charge, raising his left forearm to defend, but A's right fist crosses over the not-yet-steadied defense line on a higher arc. At about 9.00 seconds, a side medium close-up clearly shows A's right fist hitting the hard armor on the left side of B's lower jaw, with the arm still connected to A's shoulder; from the moment of real contact to 10.00 seconds, enter ultra slow motion. B's lower jaw and head first turn in the direction of force, shoulders and torso follow, the center of gravity under his feet is pushed off support, silver-white hair, steam, and fine dust lag behind; the fist surface compresses on contact and then leaves along a natural arc, not passing through the skull. The camera moves laterally a short distance along the punch path, always clearly distinguishing A's fist, B's lower jaw, and the identities of both, without irrelevant 360-degree orbiting. At 10.00 seconds immediately return to normal speed, B loses footing and staggers backward to the right, his shoulders and back first crashing into the right-side tree belt, tree trunks break after actual collision, canopies topple in the direction of force, B falls backward with the weight, shoulders and back sliding across the soil, chest facing upward. A remains on the left and immediately chases right, not admiring his own punch, not leaving a blank pause after B falls. 12.00–17.00 seconds: a three-quarter medium-wide shot on the same side of the combat axis simultaneously frames A's face and right arm, B's upper body still lying on the ground on the right, and the visible soil surface beside his body, continuously showing aiming, downward punch, dodge, and impact point. A catches up close to B, pulls back his right fist and raises it high, gaze fixed directly on B's upper chest, shoulders pressing forward, making the fist's downward path align exactly with the chest position B still occupies at that moment. As A begins smashing down toward B's chest, B is still beneath the punch path, looks up and sees the approaching right fist; only after the downward smash has already begun does B urgently roll one body-width toward the camera side, shoulders, back, and hips rolling against the ground in sequence, moving his chest out of the punch path at the last moment. A's already-committed downward smash continues along the original route, and the right fist immediately smashes into the same patch of soil that B's upper back had just been pressing, forming a fist crater, with broken roots, clods of mud, and small stones flipping outward from that impact point. The image must make it understandable that A originally intended to hit B, and only missed because B dodged in time; the fist must not find some other empty ground by A's feet, and must not begin swinging only after B has already rolled far away. The fist does not hit B. The shot first preserves the complete relationship of B's chest leaving the original position and the fist entering that position, then follows B's short lateral shift and rise: B first supports himself with his right palm and one foot, pulls the other leg back under his body, and only after both feet are steadily planted does he lift his torso, right hand leaving the ground. A's forward lean after the missed punch exposes his ribs, and B rises from a low position close in front of and slightly to the side of A, eyes locked on A's ribs and abdomen; A is still offset left, B is still offset right and closer to the camera, not circling behind A, not teleporting from lying down to standing, with the broken tree belt and fist crater preserved. 17.00–20.00 seconds: continue the close side two-person shot, B's feet bearing weight, waist and hips rotating quickly, driving a short left punch into A's exposed ribs and abdomen. After this actual contact, A's upper body folds inward, shoulders press down, long black hair whips forward, both knees bend from the force; the bodily reaction is synchronized with this punch, not falling first and then being hit. B immediately retracts his left fist, while his right shoulder sinks, right fist drawn low by his body, rear foot driving into the ground, extending knees and hips, sending the right uppercut upward from below toward A's jaw; the two actions are one continuous counterattack, without resetting the guard or stepping back to wait in between. The camera rises from a low side position with the right punch, framing B's foot support, rising torso, and A's folding upper body, emphasizing force transferring from ground through legs and hips into the fist. 20.00–23.00 seconds: at 20.00 seconds the right uppercut clearly contacts A's jaw, and from 20.00–21.00 seconds enter the second ultra slow motion segment. A's lower jaw and head are first lifted, upper chest then expands and arches back, torso moves left-back-upward along the direction transmitted by the right fist; only after the center of gravity is pushed off support do both feet briefly leave the ground, knees, legs, and black hair lagging with inertia. B remains steadily planted on the ground, right arm completing a readable uppercut arc before stopping in a natural position, not floating up with A, not holding A aloft on his fist. The camera performs one short low upward tilt and pullback, showing in the same frame the soil under B's feet and A's airborne body. At 21.00 seconds return to normal speed, A falls along a continuous low parabola toward the left-rear, immediately beginning to descend, not hanging in the air for long, not flying up into the clouds, not adding acrobatic flips. B takes a step forward with the momentum to maintain pressure, but does not add a punch in midair. 23.00–27.00 seconds: cut to a low wide shot in front of the left-side tree belt, following A's direction of motion toward the left-rear. A's upper back and shoulders first hit the tree trunk, only then does the trunk crack, and the canopy falls backward accordingly; he lands on wet soil, shoulders, back, hips, and bent legs bearing weight in sequence, sliding along the original momentum into a soil trench, pushing aside broken branches and stones before gradually stopping. The body does not bend like rubber, is not penetrated by trees; splintered wood, leaves, and clods of mud each spin and land, and the dust cloud continues advancing left along A's recent trajectory. B advances left from the right side of the clearing, with real displacement in his steps, hands still empty, stepping over previously split tree roots along the way, stopping close in front of and to the right of A, not stepping over A, not turning his back to A; B's feet and chest are already oriented toward A lying on the left, maintaining the same orientation for the subsequent roar at him. 27.00–30.00 seconds: maintain the same side low medium-wide two-person shot, A lying on the left side of the frame, B standing on the right side of the frame, both on a similar depth plane; the camera sees from the side B's left-facing facial profile, chest, and A below-left, without placing A behind B. B's feet, hips, and chest are all oriented toward A, head and neck slightly extended forward, chin lowered, both eyes continuously locked down-left on A's face and upper body, then opens his mouth toward A in a sustained intense enormous roar. The opening relies on the lower jaw dropping, ribcage vibrating with exhalation, silver-white hair and steam behind the shoulders continuously moving; during the roar the head still faces the downed opponent, not tilting up to the sky, not turning to the camera. The camera only performs an extremely slight side push-in, while preserving B's line of sight and A's location. A is still conscious, shifting from lying on his side to bracing on one forearm and the other hand, then pulling one knee under his body, beginning to raise his head to look toward B on the right, green eyes visible through flying dust; A looks back up at B, B stares down at A, forming a clear mutual gaze. A only begins to recover, does not suddenly stand up for a reversal; B continues roaring to apply pressure, without adding attacks. Both people's breathing, hair, steam, leaves, and falling dust are always in motion. Exactly at 30.00 seconds, end on the same live image of B roaring down at A and A bracing on the ground and raising his eyes preparing to recover, with no fade-out, black screen, freeze frame, or end card, leaving space for the next segment of battle. Speed and physics rules: except for the two ultra slow motion segments at 9.00–10.00 seconds for the first punch contact and 20.00–21.00 seconds for the uppercut contact, the entire piece advances with natural fast motion, and falling, chasing, side rolling, bracing up, continuous counterattack, and landing slide must have speed, not global slow motion. Each slow motion segment emphasizes only contact and the initial force, during which characters and debris still continue moving; after returning to normal speed, immediately complete the fall. Both giants generate force through ground support, and the reactions of heads, torsos, and limbs have continuous inertia; first fist contact, then body shock; first shoulders and back hit the tree, then the trunk breaks; first body or fist hits the ground, then mud and stones fly up. Preserve only the action chain of A's right horizontal punch, A's missed right downward smash, B's left punch to the ribs and abdomen, and B's right uppercut, without inserting irrelevant boxing performance, repeated punch loops, or waiting for each other. Once fallen trees, fist craters, and sliding trenches at both ends of the forest appear, they remain, and all shots maintain the same geographic relationships. No face swapping, hair color changing, armor plate transferring, extra limbs, fused hands, body penetration, clones, teleportation, energy cannons, or additional giants; no gore close-ups, dismemberment, or death display. Strictly prohibited background music: at no time in the entire piece may background music, film score, music track, melody, rhythmic percussion, chorus, song, musical sustained tone, or tonal atmospheric bed be generated. Only diegetic audio synchronized with the image and physically belonging to the scene is allowed: wind in the storm forest, thunder strike booms and natural echoes, short arc crackles, steam hisses, canopy shaking, the two giants' respective non-verbal roars and breathing, soles stepping on mud and gravel, running ground tremors, skin and biological armor movement friction, heavy dull thuds of fists hitting jaw or ribs and abdomen, fists striking the ground, tree roots breaking, shoulders and back hitting trees, tree trunks breaking, canopies falling, wood and stone landing, bodies sliding through mud, and bracing sounds. Contact slow motion may briefly elongate the real impact and air rush sounds, and when returning to normal speed, synchronously restore natural tree collision and landing sounds; do not use musical low-frequency beds. Roars must come from the corresponding giant's mouth and chest cavity and be synchronized with open-mouth exhalation, not chorus, horns, synthesizers, or score. No dialogue, no narration, no intelligible shouted words, no lyrics, no subtitles; do not use pulse sounds, orchestral hits, melodic sound effects, or rhythmic sound effects to fill the absence of music, and do not arrange footsteps and punches into drum beats. The appearance of any background music or score is judged as unqualified. music=0; no_music=1; strict_no_music=1; audio=diegetic_only; no text, character names, titles, logos, watermarks, interface, player controls, progress bars, or end cards throughout the entire piece.

A stylish young woman with short wavy brown hair, a red bomber jacket, black cargo pants, and white sneakers works alone at a colorful late-night food truck. Style: Stylized cinematic 3D animation, expressive character acting, vibrant neon lighting, exaggerated physics, detailed environments, playful comedy, quirky animated-film aesthetic. Camera: Handheld-style tracking, dramatic low angles, overhead city views, quick push-ins, wide shots, and energetic close-ups. Story: While preparing a giant taco, a strong gust of wind sends the tortilla flying across the street. She chases it through the busy neon-lit city, dodging pedestrians and scooters, until it lands on a moving delivery van. She jumps onto the van, grabs it, and triumphantly holds it up—only for her hat to fly away. She chases the hat through a neon alley, somehow keeping the tortilla balanced on her head. She finally returns to the food truck, places the tortilla down proudly, and looks at the camera. A tiny pigeon immediately steals it. Audio: Lively city ambience, footsteps, traffic, comedic whooshes, playful music, food-truck sounds, and exaggerated impact effects. No dialogue.

Create a 30-second ultra-realistic cinematic FPV destruction sequence through a dense modern city during a catastrophic collapse. The camera moves at extreme speed from ground level, weaving through traffic, falling concrete, shattered glass, dust, fire, and collapsing structures. Relentless momentum, realistic destruction physics, cinematic scale, dynamic motion blur, dramatic lighting, 4K HDR. No text, logos, subtitles, or watermarks. [0–5s] — THE COLLAPSE The FPV camera races inches above a crowded city street. Cars brake and swerve as buildings suddenly begin collapsing on both sides. Concrete fragments and glass rain down while the camera aggressively weaves between vehicles and debris. [5–10s] — NARROW ESCAPE A huge section of concrete crashes directly ahead. The camera sharply banks around it, then ducks beneath collapsing scaffolding as metal beams fall behind. Dust rapidly fills the street while the camera maintains extreme speed. [10–15s] — THE ALLEY The camera shoots into a narrow alley between towering buildings. A massive shockwave erupts behind, sending fire, dust, papers, and debris racing toward the lens. The camera accelerates harder as the alley begins collapsing around it. [15–20s] — GROUND BREAKS The asphalt suddenly splits open beneath the camera. The FPV view rolls violently as cracks race across the ground and vehicles fall sideways. The camera briefly flips almost upside down, narrowly clears the collapsing pavement, then stabilizes. [20–25s] — THROUGH THE CHAOS The camera bursts through the end of the alley into a wider boulevard engulfed in destruction. Buildings crumble in the distance, fires burn between abandoned vehicles, and enormous clouds of dust sweep across the street. [25–30s] — FINAL BURST The camera continues forward at maximum speed, weaving between overturned cars and falling debris. A final building collapses behind it, creating a gigantic wall of dust and smoke. The camera breaks through into an open intersection, revealing the full devastated skyline before rapidly pulling upward into a wide cinematic shot. STYLE: ultra-realistic FPV drone cinematography, extreme ground-level speed, physically accurate destruction, realistic concrete and glass debris, detailed smoke and fire, dynamic camera rolls, aggressive banking, natural motion blur during high-speed movement, cinematic depth of field, volumetric dust, dramatic contrast, epic blockbuster scale, 4K HDR. AUDIO: intense wind rush, engines, tires screeching, concrete impacts, collapsing metal, glass shattering, deep structural rumbles, shockwave explosions, fire, and debris rushing past the camera. No dialogue.

Browse all Seedance 2.5 prompts Explore AI Video Prompts