[CINEMATIC SETUP] Style: Professional 4K TV Sports Broadcast footage, handheld ringside camera mixed with wide crane shots. Lens/Aperture: 35mm broadcast lens, deep depth of field to keep the crowd in view. Color Grade: High-contrast, saturated "Main Event" lighting with vibrant blues and reds. Atmosphere: Volumetric haze, strobe light flashes from the crowd
作者 C Dance AI · 上滑看下一个
aesthetic: phone held up in a packed sports bar, wall of TVs glowing behind audio: crowd chant building, stomping getting louder, bass underneath timeline: - "0-4s: Phone camera inside a packed sports bar. Hundreds of people. Wall of screens behind them.
作者 C Dance AI · 上滑看下一个
{ "description": "A young player stands on a rainy urban court gripping a basketball. He explodes forward sprinting through puddles toward the hoop in the background, rain streaking past. Two fast dribbles on wet concrete. He plants his foot hard and launches into the
作者 C Dance AI · 上滑看下一个
{ "description": "Over the shoulder of a young man in a beanie and jacket on a wet urban street at golden hour. He winds up and hurls a baseball toward the building across the street. The camera stays attached to the ball as it spins forward crossing the street,
作者 C Dance AI · 上滑看下一个
{ "description": "The GoPro logo on the stormy wave begins to vibrate. The four colored squares eject outward and reshape into the limbs of a surfer holding a GoPro Hero 12 camera in one hand. The surfer drops into the barrel, arm extended filming with the camera. A
作者 C Dance AI · 上滑看下一个
A massive coastal battlefield where ocean, land, and sky collide. Three colossal creatures engage: — a sky dragon with metallic scales and storm wings — a fire colossus emerging from volcanic fissures — a giant ocean leviathan rising from the sea Opening frame: wide shot of coastline — ocean churning, volcano erupting, storm clouds gathering overhead. Hook a
作者 C Dance AI · 上滑看下一个
Create a fast-paced cinematic continuation with rapid action and dynamic camera movement. The dragon instantly launches off the cliff with explosive force — no pause. Rocks and debris scatter outward as wings snap open mid-air. Camera immediately locks into a high-speed tracking
作者 C Dance AI · 上滑看下一个
FORMAT: 15s / ONE CONTINUOUS SHOT / top-down overhead follow, no cuts SUBJECTS: A tired woman in a loose tank top and sleep shorts, slow habitual movement. Slightly smeared eyeliner, bare feet, heavy posture, detached face. ENVIRONMENT: A cramped cluttered
作者 C Dance AI · 上滑看下一个
Morning. Kitchen. (Character 1) skillfully manipulates a frying pan with multiple tentacles, cooking pancakes. Beside him, (Character 2) holds a teapot with a suction-cup tentacle, brewing fragrant coffee. All the kitchen utensils have transformed into organic shapes, and iridescent steam rises from them. (Character 3) mutters, "Mom and Dad, you look weird,"
作者 C Dance AI · 上滑看下一个
Original Dark Fantasy Action Short Film: Inside a dilapidated church, a white-clad warrior and a black-armored opponent launch their final battle amid an atmosphere like a chorus. Stained glass shatters, moonlight penetrates the smoke and dust, and benches are overturned.
作者 C Dance AI · 上滑看下一个
A speeding bus ______> A towering T-Rex mecha in the middle of a storm . Rain physics Mechanical detail. Cinematic camera work that feels straight out of IMAX .🔥
作者 C Dance AI · 上滑看下一个
Hyper-realistic biblical combat epic, IMAX cinematography, Ridley Scott scale. 11 rapid shots in 15 seconds. The impossible battle: shepherd boy vs 9-foot giant. One stone. Oil painting texture with cinematic realism. No dialogue. Percussion-driven score. Colour: bronze armour,
作者 C Dance AI · 上滑看下一个
ultra cinematic 15-second continuous action sequence, a deadly kunoichi in a sleek combat outfit stands in a snowy Japanese temple courtyard, holding a pistol in each hand, absolute control and lethal focus before instant explosive engagement ultra immersive close-proximity system, body-follow tracking + tight orbit + aggressive push-ins, camera stays extrem
作者 C Dance AI · 上滑看下一个
FORMAT: 15s / 180 BPM / ONE CONTINUOUS SHOT / 360 POV downhill stair run, viral energy, max chaos SUBJECTS: First-person cyclist, handlebars and front wheel flashing low in frame during drops and hard turns. Vendors, laundry, scooters, dogs, chickens, cars, and pedestrians erupt around the rider as sudden obstacles. ENVIRONMENT: Dense Brazilian hillside stre
作者 C Dance AI · 上滑看下一个
A female warrior with shoulder-length hair, the ends naturally flipping outward, pressed backward and slightly disheveled by air resistance during high-speed movement. She wears a dark, form-fitting tactical suit combining real fabric and worn metal elements, with
作者 C Dance AI · 上滑看下一个
Chapter 2 (0–15 seconds): Freefall · Purple AITO R7 Enters the Frame Style: Rough realism style, using real footage, 35mm handheld camera, with subtle natural shake and film grain. Camera: Single continuous third-person tracking shot (no cuts), high-speed vertical tracking throughout, no cuts. Lighting: Dazzling, high-contrast Chongqing midday sun, strong gl
作者 C Dance AI · 上滑看下一个
[Style] Dark Oriental Fantasy Transformation, cinematic texture, 8K ultra-clear, photorealistic, DaVinci high-grade color grading [Duration] 12 seconds [Scene] Bright modern home corridor (Shot 1) → Dim living room · Dark velvet sofa (Shots 2-3) [Character] Young Asian female, black long hair, Image 1 [00:00-00:01] Shot 1: Particle Dissolve. Mid-shot, bright
作者 C Dance AI · 上滑看下一个
Ultra-realistic cinematic destruction sequence, dense modern city, extreme physical danger, aggressive forward motion, handheld chase camera, collapsing architecture, debris interaction, volumetric dust, shockwaves, no stability, no safe zones, high intensity pacing, hard cuts only, no fade, no dissolve, no morph transition. Shot 1: Camera already sprinting
作者 C Dance AI · 上滑看下一个
{ "cinematography": { "camera_perspective": "extreme macro FPV tracking shot, camera tightly attached to the fairy's back, synchronized with wing flapping frequency", "lens": "8mm probe lens ultra wide macro, extreme perspective distortion, grass blades appear like towering skyscrapers at frame edges", "depth_of_field": "very shallow depth of field (f/1.4 si
作者 C Dance AI · 上滑看下一个
STYLE: Clean 2D stickman animation with thin confident lines, no shading, no textures, no dialogue, no subtitles. Only speed lines, smear frames, impact rings, skid marks, and small sharp red blood accents on the hardest hits. Brutal, stylish, fluid, old-school internet stick-fight energy. MAIN CHARACTER: A solid black stickman martial artist with a tied kun
作者 C Dance AI · 上滑看下一个
Psychic powers, monster combat, destruction, and cinematic scale, all from a single prompt
作者 C Dance AI · 上滑看下一个
A surreal Japanese scene in pink and pastel tones using Manga aesthetics. Scene composition: Bokeh background of cherry blossom trees in bloom and flowers floating away gently. Static camera with sharp focus on mid-ground where a centre-frame buff male pianist is seen from the back, dressed in a tuxedo and playing a grand piano. Frame-right is a cool looking
作者 C Dance AI · 上滑看下一个
Use @Image1 as the character reference. Keep character design, hairstyle, ears, tail, outfit, and colors perfectly consistent. 【Subject】 A blue-haired fox girl with long flowing hair, large fluffy fox ears, and a huge fluffy tail. She wears a shrine-maiden inspired outfit with loose sleeves and a short skirt. She stands in a Japanese shrine setting with a re
作者 C Dance AI · 上滑看下一个
Slow tracking shot across a minimalist office desk, steaming ceramic coffee cup releasing gentle vapor wisps, soft Monday morning sunlight streaming through sheer curtains, clean MacBook and small succulent in frame, warm beige and oak wood tones, peaceful cinematic atmosphere, 4k photorealistic, smooth camera movement
作者 C Dance AI · 上滑看下一个
浏览经过真实视频验证的 Seedance 2.0 提示词，查看生成效果，并在不离开页面的情况下使用右侧生成器继续创作。
每个案例都同时提供完整提示词和可播放视频，方便在生成前判断运镜、节奏、光线、主体一致性与视觉风格。
选择提示词后可继续修改场景，按需上传参考图，并在右侧生成器中设置比例、分辨率、时长与音频。
依次说明主体、动作、环境、镜头运动、景别、光线、视觉风格和时间节奏，保留真正会改变画面的具体信息。
它是用来描述视频场景、主体、动作、镜头、光线、节奏和视觉风格的文字指令。
可以。你可以复制提示词，也可以直接在右侧生成器中替换主体、环境、动作或运镜。
需要。上传清晰的源图片，再用提示词描述主体动作、镜头运动、光线变化和场景发展。
当前生成器最多接收 2500 个字符。超长提示词仍可用于参考，但需要先精简再生成。