Los modelos Seedance, comparados
Cinco modelos Seedance, cada uno para un trabajo distinto: tomas largas, 4K, cámara fija, series o movimiento rápido sin sonido.
Los modelos
- 9:16
- 1080p
Prompt: [Generation Goal] Generate an architectural property film as ONE SINGLE CONTINUOUS UNBROKEN TAKE. The camera flies forward at a steady walking pace from a gravel drive, in through the front door of a modern house, down a hall, across a kitchen, out through an open glass wall, and onto a terrace. There are no cuts, no dissolves, no transitions and no jumps anywhere in the video. The camera never stops moving forward and never reverses. [Subjects and Relationships] There is exactly ONE house in the entire video, and the camera stays inside its geometry for the whole take: every room the camera enters connects physically to the one before it. No people appear anywhere, indoors or outdoors, and no pets. The house is single-storey, roughly 18 metres wide across the front, clad in black-stained vertical timber with pale honed limestone at the base and a flat roof with a thin steel fascia. [Continuous Action] The take opens low, about 70 centimetres above a pale gravel drive, moving forward toward the house. Late afternoon sun comes from the left and throws the long shadows of three slender birch trees across the gravel in front of the lens. Without any cut, the camera rises smoothly to chest height, about 1.5 metres, and climbs two wide limestone steps toward a tall black pivot door standing open. The doorway fills the frame and the camera passes through it into shadow. Still in the same unbroken shot, the camera moves down a hall with wide oak floorboards, smooth lime-plaster walls and a single large unframed abstract canvas on the right. Warm light spills in from a room ahead. The camera holds a dead-straight line down the middle of the hall. Still without any cut, the camera enters an open kitchen and living space and slides past the left edge of a long island in honed dark stone. On the island sit a shallow ceramic bowl of lemons and a matte steel kettle with a thin ribbon of steam rising from the spout. One hard shaft of sun crosses the stone surface and the dust in the air is visible inside it. Still in the same take, the camera continues forward through a fully open sliding glass wall, crossing the threshold from oak floor to limestone paving. The camera moves out onto a terrace, past a low bed of ornamental grasses moving in a light wind, and along the edge of a still dark-bottomed pool that mirrors the sky. Finally the camera rises to about 2.5 metres and slows to a stop, looking out over the pool toward low hills, with the glass wall of the house visible at the left edge of frame reflecting the sunset. End state: the camera is stationary and elevated above the terrace, the pool surface is almost flat, the grasses are still moving slightly, and the frame holds steady. [Visual Treatment] High-end architectural cinematography in the register of a magazine property film. Wide-angle rectilinear lens with a horizontal field of view around 84 degrees: perspective is generous and lines stay dead straight, with no fisheye bulge and no barrel curvature on the door frames or the roof fascia. Deep focus throughout, so the gravel in the near foreground and the hills on the horizon are both legible. Late golden-hour daylight only, warm from the left outdoors, cooler and softer indoors where it bounces off plaster. Rich but restrained grade: warm limestone cream, black timber, oak honey, deep green planting, pale sky. Smooth gimbal-stabilised motion with a constant forward speed and no shake, no camera roll and no vibration. Real 24 frames per second motion. No slow-motion, no speed ramping, no lens flare, no drone shot rising over the roof, no vignette. [Maintain Consistency] One single continuous take from the first frame to the last, with no cuts, no jump cuts, no dissolves and no camera teleports. Keep the sun's direction, its colour temperature and the time of day identical from the first frame to the last. Keep the black timber cladding, the limestone, the oak flooring and the plaster walls unchanged wherever they reappear. Exactly one house, exactly one island, exactly one kettle, exactly one pool. Materials and light never reset between spaces. [Audio] (A restrained ambient neo-classical bed runs unbroken under the whole film: sustained low strings, sparse felt piano, one slow warm swell that peaks as the camera crosses onto the terrace, and no percussion at all.) <Gravel settling faintly, a light wind through birch leaves and ornamental grasses, the soft acoustic drop as the camera passes through the doorway into interior air, a faint kettle hiss, and the quiet lap of pool water at the end.> No one speaks. There is no narration and no voiceover. No lettering anywhere in the frame: no signage, no house numbers, no brand names on the kettle or fittings, no subtitles and no on-screen captions.
- 16:9
- 30s
- 480p
Prompt: [Generation Goal] Generate a candid street interview filmed entirely in first person through the camera in the interviewer's glasses, on a sunny Los Angeles street. The interviewer walks up to a man standing by a parked supercar and asks him what he does for work. The whole video is one continuous point of view from the interviewer, who is never seen. [Reference Material Roles] @Image 1 defines the man being interviewed: his facial features, short dark hair with a neat fade, trimmed stubble, oversized cream boxy tee, black tailored shorts, white sneakers, steel watch and the sunglasses pushed up on his head. It also defines the unbranded matte white supercar with its raised scissor door, the palm-lined LA street, and the harsh bright afternoon light. [Subjects and Relationships] Exactly one man is visible in the entire video, the man from @Image 1. The interviewer holds the camera and is never seen except for one hand entering the bottom of the frame. Their relationship is a stranger stopping another stranger on the street. [Event Script] 0-5 seconds: First person point of view at eye level, walking along the pavement toward the man, the frame swaying gently with each step. He is standing beside the supercar with the scissor door raised, looking down at his phone. The interviewer's hand enters the bottom of frame in a small wave. The man looks up. Off screen interviewer, natural conversational American English: {Yo, what do you do for work?} Hard cut to. 5-10 seconds: Point of view holding steady on the man from a few steps away, framed from the knees up, the car filling the right side. He pockets his phone, gives a small shrug and a half smile, relaxed and unbothered. Man, natural conversational American English: {I run marketing ads.} Off screen interviewer: {And that got you this?} Hard cut to. 10-15 seconds: Point of view tilting down the flank of the car and back up to his face as he turns his palm toward it. He nods once, eyebrows raised, matter of fact rather than boastful. Man: {Yeah. All of it.} Off screen interviewer: {What do you use?} Hard cut to. 15-20 seconds: Point of view stepping closer, the man now framed from the chest up with the open door behind him. He grins, taps his phone twice against his palm, then steps back toward the driver's seat. Man: {Kubeez. Makes every ad I run.} End state: he is lowering himself into the seat, one hand on the raised door, still looking at the camera. [Visual Treatment] Everything looks recorded through small wearable camera glasses. Eye level first person, natural walking sway and micro head movement, mild wide angle barrel distortion, deep focus so the man and the street both stay sharp. Harsh high afternoon sun, blown highlights on the pavement, strong short shadows, occasional lens flare, punchy contrasty consumer camera colour, faint digital noise. No tripod moves, no cinematic rig moves, no shallow depth of field, no colour grading. [Audio] Only the two speaking voices and the street. Every line is spoken in natural conversational American English. The man's mouth moves only on his own lines and stays naturally closed while the interviewer is speaking. <Passing traffic, distant car horn, footsteps on pavement, the low hum of the street> sit underneath. No background music of any kind. [Maintain Consistency] Keep the man's identity, face, hair and clothing from @Image 1 unchanged in every shot. Keep the car's shape, matte white colour and raised door consistent. Exactly one man is visible throughout. Keep the same street, the same harsh afternoon light and the same first person point of view across every shot. The car carries no badges, no emblems, no brand names and no lettering anywhere on its body, wheels or number plate. No subtitles, no on screen text, no captions, no lettering on shop signs, no logos on clothing, no watermark anywhere in the frame.
- 9:16
- 20s
- 480p
Prompt: [Generation Goal] Generate a five-shot cinematic fast-food commercial that builds one flame-grilled beef burger, from the raw patty hitting an open flame to the finished hero burger. Food and hands only. No faces, no packaging, no signage, no lettering. [Subjects and Relationships] There is exactly ONE beef patty and exactly ONE burger in the entire video. Never two patties, never two burgers, never a second burger in the background. The bun is a glossy sesame-seed crown, roughly 12 centimetres across, so it reads slightly wider than the gloved hand that works on it. A single cook's hands appear: the right hand holds stainless tongs and does all the lifting and placing, and the left hand stays flat on the dark stone counter at the lower left of frame, holding the board steady, and never picks anything up. The build order is fixed and never repeats an ingredient: bottom bun, flame-grilled patty, lettuce, tomato, onion, pickles, sauce, crown bun. [Event Script] 0-3 seconds: Static low tripod shot, extreme close on a black cast-iron grill grate. The raw beef patty drops onto the bars and the open gas flame flares up around it. Fat spits, the surface goes from red to brown, and dark char stripes burn across the underside. End state: the patty sits seared and steaming, flames still licking past its edge. Hard cut to. 3-6 seconds: Overhead shot looking straight down, medium, on a matte black stone counter. The tongs in the right hand lift the charred patty out of the top of frame and lay it onto the toasted bottom bun waiting in the centre. A bead of juice runs off the edge onto the bun. The left hand stays flat on the stone at the lower left. End state: the patty is centred on the bottom bun and the tongs withdraw upward out of frame. Hard cut to. 6-8 seconds: Macro insert, extreme close, camera locked at counter height so the burger fills the frame from the side. A crisp green lettuce leaf falls into place, then a thick slice of ripe tomato lands on it, water beading on the cut face. Nothing else moves. End state: lettuce and tomato are stacked on the patty, both settling. Hard cut to. 8-10 seconds: Handheld shot orbiting slowly to the left, close, at a low three-quarter angle. Two rings of white onion and four green pickle slices drop into place, then a thick ribbon of creamy sauce is drawn across the top in one pass. The frame drifts and breathes with the operator's hands. End state: the sauce ribbon lies across the pickles and the burger is one layer from finished. Hard cut to. 10-13 seconds: The camera rises smoothly from counter height to a slight overhead three-quarter, pulling out to medium-wide, and settles. The sesame crown bun descends and lands on the stack, compressing it slightly so a little sauce squeezes out at the side. Warm orange flame glow flickers on the back wall behind the counter, well out of focus. The camera pushes in a few centimetres and stops. End state: one finished flame-grilled burger stands centred and completely still on the dark stone, glossy sesame crown on top, char marks visible on the patty edge. [Visual Treatment] High-end food commercial cinematography. Hard warm key light raking across the food from the left, deep falloff into near-black behind it, so the burger and the flames are the only bright things in frame. Rich, saturated, contrasty grade: charcoal blacks, deep flame orange, ripe red tomato, bright green lettuce. Shallow depth of field on every shot, with the ingredient currently landing held sharp and everything behind it dissolving into soft bokeh. Visible steam, oil sheen on the beef, condensation on the produce. Real macro food lens character, slow deliberate camera moves, 24 frames per second motion. No slow-motion ramping, no lens flare, no drone shots. [Audio] (A cinematic advertising bed runs unbroken under the whole spot: warm analog bass, a tight kick and rim pattern at a steady mid tempo, a low brass swell that rises across the last two shots, and one bright cymbal accent as the crown bun lands.) <A wet slap as the patty meets the hot grate, fat crackling and spitting, a low whoosh of gas flame, a crisp snap of lettuce, a soft squelch of sauce, and a dull muffled thud when the crown bun lands.> No one speaks. There is no narration and no voiceover. [Maintain Consistency] Keep the counter, the black stone surface, the lighting direction and the flame glow position consistent across all five shots. Keep the bun's sesame pattern, the patty's char pattern and the ingredient order unchanged once established. The right hand always holds the tongs and the left hand always stays flat on the counter. Exactly one patty and exactly one burger throughout. No lettering anywhere in the frame: no signage, no printed wrappers, no branded packaging, no logos or emblems on the tongs, gloves or counter, no subtitles and no on-screen captions.
- 16:9
- 13s
- 480p
Seedance 2.5
Hasta 30 segundos en una sola toma continua.
Una generación dura de 4 a 30 segundos, así que una pieza corta completa (gancho, desarrollo, remate) sale en una sola toma, sin cortes, y un borrador a 480p te deja revisarla antes del final en 1080p.
Ideal para
- Piezas completas de 15 a 30 s
- Producto, personaje y voz en un render
- Diálogo con guion, español incluido
Mejor otro modelo para
- Entregas en 4K. Usa Seedance 2 Standard.
- Diálogo en idiomas fuera de la lista oficial.
Especificaciones
- Modos
- Texto o primer fotograma, último opcional
- Duración
- De 4 a 30 segundos
- Resolución
- 480p, 720p, 1080p. Sin 4K
- Referencias
- 30 imágenes, 10 vídeos, 10 audios
- Audio
- Nativo, activado por defecto
- Diálogo
- 11 idiomas con soporte oficial
Ver preciosOcultar precios
Precio
| Opción | Tarifa | Unidad |
|---|---|---|
| 480p | 29 | créditos / s |
| 480p, con vídeo de referencia | 17 | créditos / s |
| 720p | 61 | créditos / s |
| 720p, con vídeo de referencia | 36 | créditos / s |
| 1080p | 142 | créditos / s |
| 1080p, con vídeo de referencia | 85 | créditos / s |
Prompt: [Generation Goal] Generate a five-shot clip of a packed rooftop night party high above a city, built around one woman in the middle of the crowd. Real event footage energy with hard cuts. [Subjects and Relationships] Exactly ONE hero woman appears and she is the only person ever held in sharp focus: mid twenties, long box braids gathered high with the bottom third of every braid dyed electric blue, silver hoops, cropped black tank under a loose mesh overshirt, black cargo trousers. She holds a plain matte black drinks can in her right hand the whole clip. Around her roughly forty people dance shoulder to shoulder, always softer than she is, never a second subject, no face among them legible. A scaffold light rig stands at the far end of the roof, its operator only ever a black silhouette. A skyline of dark towers with lit windows sits behind the parapet throughout. [Event Script] 0-3 seconds: High wide shot from a roof corner looking down across the crowd, drifting slowly right. Bodies packed to the parapet, four hard magenta and cyan beams sweeping through thick haze, confetti still falling from an earlier burst. She is a small figure near the centre with one beam catching her braids. End state: the beams cross and swing back, the crowd still moving. Hard cut to. 3-6 seconds: Handheld medium shot from inside the crowd at chest height, framed on her from the waist up, drifting as the operator is jostled. Silhouetted shoulders and arms wipe the near foreground. She dances in place, lifts the can, sips, lowers it and looks off toward the rig. End state: the can rests at her chest and confetti settles on her shoulder. Hard cut to. 6-8 seconds: Macro insert locked close on the can, the crowd behind it dissolved into round coloured bokeh. Condensation runs down the metal, a strip of confetti falls past, one cyan beam rakes across. End state: the can is still and wet. Hard cut to. 8-11 seconds: Extreme low shot from between people at hip height, tilted steeply up past raised hands into the night sky, rotating slowly clockwise. Forearms and open palms fill the lower frame, the haze glows where the beams cut it, confetti tumbles toward the lens. End state: a fresh confetti burst opens overhead and falls. Hard cut to. 11-15 seconds: Slow push in from the front at eye level, medium to close, the crowd blurring out around her. She turns to the lens, breaks into a wide open smile and raises the can toward camera in a small toast. End state: she is centred and close, can raised beside her cheek, confetti drifting through frame. [Visual Treatment] Handheld night event photography, honest festival recap rather than polished commercial. Wide rectilinear lenses near 84 degrees, no fisheye bulge, macro optics on the insert. No light source but the rig: saturated magenta, cyan and deep orange beams cutting heavy haze, coloured rim light on shoulders and braids, deep unlit black between bodies. Contrasty night grade, crushed blacks, high ISO grain, shallow focus on the hero shots. Real 24 frames per second motion, confetti tumbling on real air resistance. No slow-motion, no speed ramping, no drone shot, no graphic overlays, no drawn light trails. [Audio] (A driving instrumental house bed runs unbroken: four on the floor kick, warm analog sub bass, filtered synth stabs, open hi-hats, one rising sweep on the final push in. No vocals.) <A dense crowd cheering, hands clapping on the beat, a compressed air pop as the confetti cannon fires, wind across an open roof.> No one speaks. [Maintain Consistency] Keep her identity, braids and blue ends, clothing and the can unchanged in all five shots. Keep the roof, the skyline, the rig and the beam colours identical. Exactly one hero woman, exactly one can, exactly one light rig. The can stays completely blank: no logo, no emblem, no brand name, no lettering. No lettering anywhere else in frame: no signage, no banners, no text on clothing, no subtitles, no captions.
- 16:9
- 15s
- 720p
Prompt: [Generation Goal] Generate a five-shot action sequence: one red mid-engine supercar runs an alpine pass in Switzerland while a grey twin-tail fighter jet attacks it from the air. Blockbuster chase energy, real crash physics. [Subjects and Relationships] Exactly ONE supercar: a low wedge-profile mid-engine two seater in deep gloss red, black five-spoke wheels, carbon front splitter, low fixed rear wing, four round tail lights, dark glass, no visible driver. Exactly ONE aircraft: a grey twin-engine fighter with two canted tails, bubble canopy and square intakes, its pilot only a helmeted silhouette. The jet always comes from behind or above, the car always runs forward. No other vehicle and no pedestrian appears. The road is a two lane alpine pass with a low stone parapet, cut into a green mountainside above a deep blue lake, snow capped peaks opposite, and one short rock tunnel. [Event Script] 0-3 seconds: Low tracking shot from the front at bumper height, camera running backwards ahead of the car as it comes hard out of a right hand bend, heat shimmer off the tarmac, grit off the verge. Its shadow darkens as the jet crosses the sun overhead. End state: the car keeps closing on the lens. Hard cut to. 3-6 seconds: Wide side shot across the valley, the pass cutting the frame diagonally. The jet drops behind the car and levels off metres above the tarmac, its jetwash tearing dust and pine needles off the roadside. End state: the jet sits directly behind the car at matching speed. Hard cut to. 6-9 seconds: Low front three quarter shot from outside the bend. Cannon rounds walk up the road in a line of bursting tarmac, the parapet blows apart into flying stone, the car swerves left through it. A stone block strikes the front wing and tears the splitter off. End state: the car is sideways, tail out, smoke off the rear tyres. Hard cut to. 9-12 seconds: Camera locked low inside the tunnel mouth looking out. The car catches the drift, straightens and rockets in toward the lens. The jet pulls up hard short of the rock face, afterburner flaring white blue, dust filling the tunnel mouth behind the car. End state: the car fills the frame in near darkness, tail lights burning. Hard cut to. 12-15 seconds: Wide low shot at the far tunnel exit as the car bursts into hard sunlight, damaged front end trailing thin smoke, and slows. The jet climbs away vertically over the lake and shrinks to a point. End state: the car is stationary, dust settling, the sky empty. [Visual Treatment] Big budget action cinematography. Long telephoto compression on the chase shots so the mountain stacks flat behind the car, wide rectilinear lenses elsewhere. Hard alpine sunlight, deep saturated colour, clean blue sky, cold shadow in the rock cuts. Real 24 frames per second motion with heavy natural blur and handheld weight on the chase shots. No slow-motion, no speed ramping, no lens flare, no drone shot, no graphic overlays, no CGI render finish. The car squats under power and steps the rear out on the swerve, debris flies on ballistic paths and falls under gravity, dust rolls in the jetwash rather than puffing on the spot, and the jet needs distance to pull up. [Audio] (No music.) <A high revving engine climbing through the gears with overrun crackle, tyre roar and a long scrub as the tail steps out, stone shattering, debris raining onto tarmac, and the tearing roar of a low jet with afterburner rumble.> No one speaks. [Maintain Consistency] Keep the car's shape, red paint, black wheels and damage continuity: once the splitter is torn off it stays off. Keep the jet's grey paint and shape identical. Keep the pass, the lake and the sun direction the same in every shot. Exactly one car and exactly one aircraft. No badge, no crest, no emblem, no marque, no wordmark and no number plate text on the car. No roundel, no insignia and no tail code on the aircraft. No lettering anywhere in frame: no road sign text, no subtitles, no captions.
- 16:9
- 15s
- 720p
Prompt: Raccoon trying desperately to open a fridge in a suburban kitchen at night, tiny frustrated grunts and metallic fridge clanks, deadpan lighting, viral sitcom vibe, 16:9 landscape, 5 seconds.
- 16:9
- 5s
- 480p
Seedance 2
El único Seedance que renderiza en 4K.
Seedance 2 es la opción cuando el plano tiene que ser grande. Su nivel Standard renderiza hasta 4K, para planos principales que vas a recortar o reencuadrar en postproducción, y el nivel Fast de la misma tarjeta hace borradores rápidos.
Ideal para
- Planos principales en 4K para reencuadrar
- Transformaciones entre dos fotogramas exactos
- Borradores en Fast, final en Standard
Mejor otro modelo para
- Tomas de más de 15 segundos. Usa Seedance 2.5.
- Una sola imagen sin fotograma final en image-to-video.
Especificaciones
- Modos
- Texto o primer y último fotograma
- Niveles
- Standard y Fast, misma tarjeta
- Duración
- De 4 a 15 segundos
- Resolución
- Standard hasta 4K, Fast 720p
- Referencias
- 9 imágenes, 3 vídeos, 3 audios
- Audio
- Nativo, activado por defecto
Ver preciosOcultar precios
Precio
| Opción | Tarifa | Unidad |
|---|---|---|
| Estándar, 480p | 22 | créditos / s |
| Estándar, 480p, con vídeo de referencia | 17 | créditos / s |
| Estándar, 720p | 39 | créditos / s |
| Estándar, 720p, con vídeo de referencia | 29 | créditos / s |
| Estándar, 1080p | 87 | créditos / s |
| Estándar, 1080p, con vídeo de referencia | 55 | créditos / s |
| Estándar, 4K | 260 | créditos / s |
| Estándar, 4K, con vídeo de referencia | 165 | créditos / s |
| Fast, 480p | 19 | créditos / s |
| Fast, 480p, con vídeo de referencia | 15 | créditos / s |
| Fast, 720p | 31 | créditos / s |
| Fast, 720p, con vídeo de referencia | 22 | créditos / s |
Prompt: Extreme macro of a monarch butterfly wing against a deep black studio backdrop. The frame opens heavily soft and out of focus, then resolves continuously and evenly into razor-sharp detail: individual overlapping scales, iridescent orange and black edges, fine dust motes catching the light. One slow steady push-in, no cuts, single controlled key light from the upper left, rich blacks. Completely blank of any writing, lettering, captions, logos or user interface.
- 4:3
- 4s
- 720p
Prompt: A brass pocket watch on dark velvet, shot with a deliberate camera move: a slow dolly in, then a smooth arc around the watch face, then a controlled crane up looking down on it. Precise, mechanical camera motion, the subject dead centre the whole time. Single soft key light, one continuous shot. Completely blank of any writing, lettering, captions, logos or user interface.
- 1:1
- 4s
- 720p
Prompt: A matte amber glass skincare bottle rotating slowly on a polished stone plinth in a photography studio. Soft light sweeps across it, a second rim light kicks in from behind, water droplets bead on the glass. Deep charcoal seamless backdrop, commercial product photography, one continuous shot. Completely blank of any writing, lettering, captions, logos or user interface.
- 1:1
- 4s
- 720p
Seedance 2 Mini
Las entradas de Seedance 2, para series, hasta 720p.
Seedance 2 Mini admite lo mismo que Seedance 2, también de 4 a 15 segundos, pero solo renderiza a 480p o 720p. Encaja en catálogos: una foto por referencia, una plantilla de prompt, decenas de clips verticales.
Ideal para
- Un clip para cada producto del catálogo
- Tandas de clips verticales para redes
- Probar referencias antes de un render Standard
Mejor otro modelo para
- 1080p o 4K. Usa Seedance 2 Standard.
- Clips de más de 15 segundos. Usa Seedance 2.5.
Especificaciones
- Modos
- Texto o primer y último fotograma
- Duración
- De 4 a 15 segundos
- Resolución
- 480p, 720p
- Referencias
- 9 imágenes, 3 vídeos, 3 audios
- Audio
- Nativo, activado por defecto
Ver preciosOcultar precios
Precio
| Opción | Tarifa | Unidad |
|---|---|---|
| 480p | 12 | créditos / s |
| 480p, con vídeo de referencia | 8 | créditos / s |
| 720p | 24 | créditos / s |
| 720p, con vídeo de referencia | 15 | créditos / s |
Prompt: Vertical cinematic ad, slow reveal of hot espresso being poured into a ceramic cup, warm ambient barista sounds and gentle piano, golden morning light, 9:16 vertical.
- 9:16
- 4s
- 720p
Prompt: Close-up photoreal shot of a bearded man in a denim jacket laughing and looking at the camera, natural window light, handheld steady, shallow depth of field, quiet indoor ambience, 9:16 vertical.
- 9:16
- 4s
- 720p
Prompt: Fashion model in flowing silk dress spinning slowly in a minimalist white studio, fabric catching soft overhead light, dreamy slow-motion movement, editorial framing, 9:16 vertical.
- 9:16
- 4s
- 720p
Seedance 1.5 Pro
Clips de 4, 8 o 12 segundos con cámara fija.
Seedance 1.5 Pro es el único Seedance que puede fijar la cámara. Activas la lente fija y el encuadre se queda quieto mientras el sujeto se mueve: un producto como en trípode, un cinemagraph, un fondo en bucle.
Ideal para
- Planos fijos de producto y bodegón
- Transiciones de inicio a final en 1080p
- Decidir en cada generación si hay sonido
Mejor otro modelo para
- Referencias en vídeo o audio: Seedance 2 o 2.5.
- Duraciones distintas de 4, 8 o 12 segundos.
Especificaciones
- Modos
- Texto o fotograma inicial y final
- Duración
- 4, 8 o 12 segundos
- Resolución
- 480p, 720p, 1080p
- Relación de aspecto
- Seis relaciones, sin adaptativa
- Cámara
- Interruptor de lente fija
- Audio
- Opcional en cada generación
Ver preciosOcultar precios
Precio
| Opción | Tarifa | Unidad |
|---|---|---|
| 480p, 4 s | 12 | créditos |
| 480p, 4 s, con audio | 24 | créditos |
| 480p, 8 s | 24 | créditos |
| 480p, 8 s, con audio | 48 | créditos |
| 480p, 12 s | 48 | créditos |
| 480p, 12 s, con audio | 58 | créditos |
| 720p, 4 s | 19 | créditos |
| 720p, 4 s, con audio | 38 | créditos |
| 720p, 8 s | 38 | créditos |
| 720p, 8 s, con audio | 76 | créditos |
| 720p, 12 s | 76 | créditos |
| 720p, 12 s, con audio | 152 | créditos |
| 1080p, 4 s | 39 | créditos |
| 1080p, 4 s, con audio | 78 | créditos |
| 1080p, 8 s | 78 | créditos |
| 1080p, 8 s, con audio | 156 | créditos |
| 1080p, 12 s | 156 | créditos |
| 1080p, 12 s, con audio | 312 | créditos |
Prompt: The scene begins completely still, as if it were just a normal painting hanging in the gallery. Soft museum spotlights illuminate the frame and create gentle shadows on the wall. The camera slowly dollies forward toward the painting, moving smoothly like a documentary camera observing artwork in a gallery. After a few seconds, the cat inside the painting subtly moves its head and eyes, breaking the stillness of the portrait. The cat then reaches one paw out of the painting frame, transitioning from painted brushstroke texture into fully realistic fur and lighting. The cat climbs out of the painting frame onto the wall area, becoming fully three-dimensional and realistic. The camera continues moving forward as if following the subject, maintaining focus on the cat as it approaches. The empty painting remains hanging behind on the wall. Cinematic lighting, realistic motion, seamless transition from oil painting texture to photorealistic fur, documentary-style camera movement, ultra detailed.
- 1:1
- 5s
Seedance 1.0
Una imagen, un clip mudo de 5 o 10 segundos.
Seedance 1.0 hace una sola cosa: anima una imagen. Sin text-to-video, sin referencias, sin fotograma final, sin control de aspecto. Subes una foto, describes el movimiento y eliges duración y resolución. El resultado siempre es mudo.
Ideal para
- Movimiento en una foto que ya tienes
- Bucles mudos para webs y presentaciones
Mejor otro modelo para
- Cualquier clip que necesite sonido generado.
- Empezar desde texto o guiar con referencias.
Especificaciones
- Modos
- Solo image-to-video, una imagen
- Duración
- 5 o 10 segundos
- Resolución
- 720p, 1080p
- Entrada
- JPEG, PNG o WebP, 10 MB
- Audio
- Ninguno, siempre mudo
Ver preciosOcultar precios
Precio
| Opción | Tarifa | Unidad |
|---|---|---|
| 720p, 5 s | 40 | créditos |
| 720p, 10 s | 80 | créditos |
| 1080p, 5 s | 80 | créditos |
| 1080p, 10 s | 130 | créditos |
Los modelos Seedance frente a frente
| Modelo | Duración | Resolución | Audio | Referencias | Solo aquí |
|---|---|---|---|---|---|
 Seedance 2.5 | 4 a 30 s | Hasta 1080p | Activado, 11 idiomas | Hasta 50 | Borrador a 480p |
 Seedance 2 | 4 a 15 s | Hasta 4K | Activado | Hasta 15 | Salida en 4K |
 Seedance 2 Mini | 4 a 15 s | Hasta 720p | Activado | Hasta 15 | Pensado para series |
 Seedance 1.5 Pro | 4, 8, 12 s | Hasta 1080p | Opcional | Una imagen opcional | Cámara fija |
 Seedance 1.0 | 5 o 10 s | 720p o 1080p | Mudo | Ninguna | Una imagen, nada más |
¿Qué Seedance me conviene?
Una toma de más de 15 segundos
Seedance 2.5
El único Seedance que renderiza hasta 30 segundos en una toma.
Un máster en 4K
Seedance 2
Standard es el único nivel Seedance con 4K.
Muchos clips de producto o redes
Seedance 2 Mini
Volumen sin renunciar a las referencias.
Una transformación exacta entre dos fotogramas
Seedance 2
Image-to-video siempre pide primer y último fotograma, hasta 4K.
Cámara quieta, solo el sujeto se mueve
Seedance 1.5 Pro
La lente fija congela el encuadre. Ningún otro Seedance la tiene.
Movimiento rápido y mudo sobre una foto
Seedance 1.0
Una imagen de entrada, 5 o 10 segundos, nada que preparar.
Cómo encaja cada modelo
Seedance 2.5 renderiza tomas continuas de hasta 30 segundos. Seedance 2 es el único con 4K, y Seedance 2 Mini admite las mismas entradas, hasta 720p. Seedance 1.5 Pro añade la cámara fija. Seedance 1.0 anima una sola imagen, sin sonido.
Cómo sacar partido a Seedance
- En Seedance 2 y Mini, para animar una sola imagen, usa text-to-video y adjúntala como referencia.
- En Seedance 2.5, escribe las tomas largas en bloques con marcas de tiempo para ordenar la acción.
- Cita los archivos como @Image 1, @Video 1 o @Audio 1, por orden de subida en cada tipo.
Elige un plan
Un solo saldo de créditos cubre imagen, vídeo y audio. Elige el tamaño que encaje con lo que produces.
Studio Pro
Valor de estudio con mayor margen.
2 puestos incluidos
Lo que generarás
- Imágenes~1600
- Canciones~492
- Vídeos~640
- Separaciones~1066
- Minutos de subtítulos~213
- Todos los modelosIncluido
Starter
Generaciones de nivel básico para uso regular.
Lo que generarás
- Imágenes~275
- Canciones~84
- Vídeos~110
- Separaciones~183
- Minutos de subtítulos~36
- Todos los modelosIncluido
Pro
Accede a la experiencia completa.
Lo que generarás
- Imágenes~650
- Canciones~200
- Vídeos~260
- Separaciones~433
- Minutos de subtítulos~86
- Todos los modelosIncluido
Powerhouse
La opción más rentable.
13.000 créditos/mes
Puestos extra 19 €/puesto/mes
Lo que generarás
- Imágenes~3250
- Canciones~1000
- Vídeos~1300
- Separaciones~2166
- Minutos de subtítulos~433
- Todos los modelosIncluido
Enterprise
Para equipos que superan los planes estándar.
Precios adaptados a tu equipo
Qué incluye
- Número de puestos personalizado
- Volumen de créditos personalizado
- Pago por factura
- Soporte prioritario
Starter
Generaciones de nivel básico para uso regular.
Lo que generarás
- Imágenes~275
- Canciones~84
- Vídeos~110
- Separaciones~183
- Minutos de subtítulos~36
- Todos los modelosIncluido
Pro
Accede a la experiencia completa.
Lo que generarás
- Imágenes~650
- Canciones~200
- Vídeos~260
- Separaciones~433
- Minutos de subtítulos~86
- Todos los modelosIncluido
Studio Pro
Valor de estudio con mayor margen.
2 puestos incluidos
Lo que generarás
- Imágenes~1600
- Canciones~492
- Vídeos~640
- Separaciones~1066
- Minutos de subtítulos~213
- Todos los modelosIncluido
Powerhouse
La opción más rentable.
13.000 créditos/mes
Puestos extra 19 €/puesto/mes
Lo que generarás
- Imágenes~3250
- Canciones~1000
- Vídeos~1300
- Separaciones~2166
- Minutos de subtítulos~433
- Todos los modelosIncluido
Enterprise
Para equipos que superan los planes estándar.
Precios adaptados a tu equipo
Qué incluye
- Número de puestos personalizado
- Volumen de créditos personalizado
- Pago por factura
- Soporte prioritario
Prueba cualquier modelo con packs de créditos de un solo pago
Los créditos nunca caducan y no requieren suscripción — compra un pack, genera, comprueba si Kubeez encaja contigo.
Recarga rápida
Recarga rápida para ráfagas cortas de generación.
Créditos
500
Pagas
6 €
Pack estándar
Pack mediano al tipo estándar por crédito.
Créditos
1500
Pagas
18 €
Pack de volumen
Pack mayor con descuento por volumen incluido.
Créditos
5000
Pagas
60 €57 €
Pack de producción
Mejor precio por crédito para producción sostenida.
Créditos
15.000
Pagas
180 €162 €
Preguntas frecuentes
No, es un modelo distinto. Admite más referencias, acepta audio de referencia por sí solo, solo necesita el primer fotograma para image-to-video y tiene modo borrador. Seedance 2 conserva el 4K.
Sí, en Seedance 2.5, con 11 idiomas oficiales: chino, inglés, español, indonesio, malayo, tailandés, árabe, portugués, vietnamita, japonés y coreano. Seedance 2 no publica lista. Para otros, añade una locución con el audio de Kubeez.
Seedance 2, 2 Mini y 2.5 rechazan imágenes y vídeos que puedan mostrar a una persona real, también personas fotorrealistas generadas. Usa un producto, lugar o animal, o describe a la persona en el prompt.
Image-to-video fija fotogramas exactos: en Seedance 2 y Mini, primer y último fotograma obligatorios. Las referencias dicen qué debe aparecer, un producto, un estilo, una voz, sin fijar ningún fotograma.
Promocionas un borrador a 480p a un final en 1080p en 7 días como máximo. El final reutiliza su prompt, referencias, duración y encuadre, así que apruebas la toma, no la cambias.
Abre Seedance en el estudio
Elige el modelo que encaja con tu trabajo, escribe tu prompt y genera.