Seedance 2.5 Text to Video API

bytedance/seedance-v2.5/text-to-video

Metin prompt’undan senkronize sesli video üretin — 480p ya da 720p’de, 30 saniyeye kadar uzayan tek çekimler.

Çözünürlüğe göre fiyatlandırılır

OpenAPI

Input

Prompt

One continuous take; the camera never accelerates abruptly, its speed stays even.
16:9 cinematic widescreen | 8K ultra HD | UE5 photoreal CG
Late-night CBD rooftop, elite-ensemble version | closing helicopter wide shot

Global parameters:
16:9 cinematic ultra-wide, 8K ultra HD, cinema-grade photoreal CG, UE5 engine, extreme PBR physical materials, ultra-sharp detail, filmic grain. Late-night open-air rooftop of a CBD office tower, a colossal city nightscape on every side, towering neon streaks, ribbons of traffic light in the distance. Metal railings and cold-white industrial lamps along the roof edge; overall a cold-blue low-key night, high-contrast low-light atmosphere, city backlight silhouettes, low saturation and high contrast for suspense and pressure. Ambient sound throughout: high-altitude night wind howling, the low-frequency rumble of distant city traffic, fabric rustle, the bass of a helicopter rotor far off. No BGM, no subtitles, no watermark.

Shot 1 | 00:00-00:02 Medium opening, arcing left
Open on a medium shot; after a steady start the camera arcs slowly to the left. Night wind surges across the rooftop. A 22-year-old woman with supermodel presence, in a black suit, stands side-on at the roof edge, head lowered in thought, then lifts her eyes haughtily straight into the lens — distant, cold, disdainful. The foreground falls slightly out of focus and leaves negative space; focus locks hard on her. Deep layering in the night view, city neon rim-lighting her shoulders and neck, a dark-toned high-end cinematic feel.

Shot 2 | 00:02-00:05 Rack focus, extreme close-up on the first man
The camera keeps drifting left while racking focus rapidly to the foreground. The blurred foreground snaps sharp: a 22-year-old man in a black suit and fine gold-rimmed glasses whips his head round; an extreme low-angle close-up fills the foreground. His lenses reflect the city's cold neon across their full width, his gaze cold and contemptuous, the pressure intense. The woman behind recedes into total blur — an extreme push-pull between foreground and background, low light, high contrast, suspense at maximum.

Shot 3 | 00:05-00:09 Pull back and pan, the second woman seizes the frame
The camera pulls back fast while continuing to pan left. The bespectacled man recedes to a medium shot, arms hanging naturally, standing tall. From the left edge a second 22-year-old woman cuts in — sharp-featured, in a bold white suit — seizing centre frame in close-up. One hand tugs lightly at her lapel, chin raised, a mocking cold smile at the corner of her mouth, looking down on the lens from above; her presence razor-sharp and domineering. Focus locks precisely on her face, the city nightscape streaming behind her as layered background.

Shot 4 | 00:09-00:12 Diagonal retreat and crane up, the second man's entrance
The camera retreats diagonally across a wide range while craning upward, the field of view opening rapidly. Close on the figure at dead centre for the final entrance: a 22-year-old man with the presence of someone in total control, white suit open, lounging side-on against the rooftop's metal railing, both hands spread naturally along it, posture relaxed and commanding, gaze indifferent, in command of everything, carrying the weight of real power. The neon city night fills the far background, light and shadow at extreme contrast.

Shot 5 | 00:12-00:15 Wide closing frame, the four-way power portrait

The camera pulls out to eye level; all four share the frame.

Describe the scene, the action, the camera move, and the sound you want in the clip.

resolution

Output resolution of the clip. 720p costs more per second of video than 480p.

aspect_ratio

Frame shape of the clip. Use adaptive to let the model pick the framing.

Duration

Clip length in seconds, from 4 to 30. Leave at -1 (the default) to let the model choose an appropriate whole-second length inside that range. Billing is per second of finished video, so -1 makes the cost of a run variable.

Additional Settings

Customize your input with more control.

Generate Audio

Generate a synchronized soundtrack (dialogue, ambience and sound effects) together with the picture. Set false for a silent clip; the price is the same either way.

You need to be logged in to run this model and view results.

Output

{
  "error": "",
  "inferenceTime": 0,
  "output": "https://media.modelrunner.ai/qDfZnddgokFjmas1-cbd-rooftop-power-ensemble-output.webm",
  "input": {
    "prompt": "One continuous take; the camera never accelerates abruptly, its speed stays even.\n16:9 cinematic widescreen | 8K ultra HD | UE5 photoreal CG\nLate-night CBD rooftop, elite-ensemble version | closing helicopter wide shot\n\nGlobal parameters:\n16:9 cinematic ultra-wide, 8K ultra HD, cinema-grade photoreal CG, UE5 engine, extreme PBR physical materials, ultra-sharp detail, filmic grain. Late-night open-air rooftop of a CBD office tower, a colossal city nightscape on every side, towering neon streaks, ribbons of traffic light in the distance. Metal railings and cold-white industrial lamps along the roof edge; overall a cold-blue low-key night, high-contrast low-light atmosphere, city backlight silhouettes, low saturation and high contrast for suspense and pressure. Ambient sound throughout: high-altitude night wind howling, the low-frequency rumble of distant city traffic, fabric rustle, the bass of a helicopter rotor far off. No BGM, no subtitles, no watermark.\n\nShot 1 | 00:00-00:02 Medium opening, arcing left\nOpen on a medium shot; after a steady start the camera arcs slowly to the left. Night wind surges across the rooftop. A 22-year-old woman with supermodel presence, in a black suit, stands side-on at the roof edge, head lowered in thought, then lifts her eyes haughtily straight into the lens — distant, cold, disdainful. The foreground falls slightly out of focus and leaves negative space; focus locks hard on her. Deep layering in the night view, city neon rim-lighting her shoulders and neck, a dark-toned high-end cinematic feel.\n\nShot 2 | 00:02-00:05 Rack focus, extreme close-up on the first man\nThe camera keeps drifting left while racking focus rapidly to the foreground. The blurred foreground snaps sharp: a 22-year-old man in a black suit and fine gold-rimmed glasses whips his head round; an extreme low-angle close-up fills the foreground. His lenses reflect the city's cold neon across their full width, his gaze cold and contemptuous, the pressure intense. The woman behind recedes into total blur — an extreme push-pull between foreground and background, low light, high contrast, suspense at maximum.\n\nShot 3 | 00:05-00:09 Pull back and pan, the second woman seizes the frame\nThe camera pulls back fast while continuing to pan left. The bespectacled man recedes to a medium shot, arms hanging naturally, standing tall. From the left edge a second 22-year-old woman cuts in — sharp-featured, in a bold white suit — seizing centre frame in close-up. One hand tugs lightly at her lapel, chin raised, a mocking cold smile at the corner of her mouth, looking down on the lens from above; her presence razor-sharp and domineering. Focus locks precisely on her face, the city nightscape streaming behind her as layered background.\n\nShot 4 | 00:09-00:12 Diagonal retreat and crane up, the second man's entrance\nThe camera retreats diagonally across a wide range while craning upward, the field of view opening rapidly. Close on the figure at dead centre for the final entrance: a 22-year-old man with the presence of someone in total control, white suit open, lounging side-on against the rooftop's metal railing, both hands spread naturally along it, posture relaxed and commanding, gaze indifferent, in command of everything, carrying the weight of real power. The neon city night fills the far background, light and shadow at extreme contrast.\n\nShot 5 | 00:12-00:15 Wide closing frame, the four-way power portrait\n\nThe camera pulls out to eye level; all four share the frame.",
    "duration": 15,
    "resolution": "480p",
    "aspect_ratio": "9:16",
    "generate_audio": true
  },
  "logs": ""
}

Logs (1 lines)

Örnekler

Fiyatlandırma

Fiyat, çıktı videonuzun hedef çözünürlüğüne göre değişir.

480p

$0.154

çıktı videosunun saniyesi başına

yani yaklaşık $1 ile 6 saniye

720p

$0.347

çıktı videosunun saniyesi başına

yani yaklaşık $1 ile 3 saniye

Seedance 2.5 Text to Video API

Seedance 2.5 Text to Video is a text-to-video AI model by bytedance. On ModelRunner it runs through a REST API or via MCP from any AI assistant, at $0.347 per second of video.

POST https://queue.modelrunner.run/bytedance/seedance-v2.5/text-to-video

cURL

# Submit a request to the queue. Input fields go at the top level of the
# body. The optional reserved "metadata" object holds your own flat string
# tags — stored on the request, never sent to the model; filter later with
# GET https://queue.modelrunner.run/requests?metadata=<url-encoded JSON>.
curl -X POST https://queue.modelrunner.run/bytedance/seedance-v2.5/text-to-video \
  -H "Authorization: Key $MRUN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "One continuous take; the camera never accelerates abruptly, its speed stays even.\n16:9 cinematic widescreen | 8K ultr…",
    "duration": 15,
    "resolution": "480p",
    "aspect_ratio": "9:16",
    "generate_audio": true,
    "metadata": {
      "project": "my-project"
    }
  }'
# → { "request_id": "...", "status_url": "...", "response_url": "..." }

# Poll status_url until "COMPLETED", then fetch the result
curl "https://queue.modelrunner.run/bytedance/seedance-v2.5/text-to-video/requests/$REQUEST_ID/status" \
  -H "Authorization: Key $MRUN_API_KEY"
curl "https://queue.modelrunner.run/bytedance/seedance-v2.5/text-to-video/requests/$REQUEST_ID" \
  -H "Authorization: Key $MRUN_API_KEY"

JavaScript

import { modelrunner } from "@modelrunner/client";

const result = await modelrunner.subscribe("bytedance/seedance-v2.5/text-to-video", {
  input: {
    "prompt": "One continuous take; the camera never accelerates abruptly, its speed stays even.\n16:9 cinematic widescreen | 8K ultr…",
    "duration": 15,
    "resolution": "480p",
    "aspect_ratio": "9:16",
    "generate_audio": true
  },
});
console.log(result);

Python

import os
import requests

headers = {"Authorization": f"Key {os.environ['MRUN_API_KEY']}"}

submitted = requests.post(
    "https://queue.modelrunner.run/bytedance/seedance-v2.5/text-to-video",
    headers=headers,
    json={
      "prompt": "One continuous take; the camera never accelerates abruptly, its speed stays even.\n16:9 cinematic widescreen | 8K ultr…",
      "duration": 15,
      "resolution": "480p",
      "aspect_ratio": "9:16",
      "generate_audio": true
    },
).json()

# Poll submitted["status_url"] until "COMPLETED", then:
result = requests.get(submitted["response_url"], headers=headers).json()

Input parameters

Input parameters of Seedance 2.5 Text to Video
Name	Type	Required	Description
prompt	string	yes	Describe the scene, the action, the camera move, and the sound you want in the clip.
resolution	enum	no	Output resolution of the clip. 720p costs more per second of video than 480p. One of: `480p`, `720p`. Default: "720p".
aspect_ratio	enum	no	Frame shape of the clip. Use adaptive to let the model pick the framing. One of: `16:9`, `4:3`, `1:1`, `3:4`, `9:16`, `21:9`, `adaptive`. Default: "16:9".
duration	integer	no	Clip length in seconds, from 4 to 30. Leave at -1 (the default) to let the model choose an appropriate whole-second length inside that range. Billing is per second of finished video, so -1 makes the cost of a run variable. Default: -1.
generate_audio	boolean	no	Generate a synchronized soundtrack (dialogue, ambience and sound effects) together with the picture. Set false for a silent clip; the price is the same either way. Default: true.

Machine-readable: OpenAPI schema · llms.txt

Use Seedance 2.5 Text to Video from Claude & Cursor (MCP)

Point Claude Code, Claude Desktop, Cursor, or any MCP client at the ModelRunner MCP server and Seedance 2.5 Text to Video becomes a tool your assistant can call directly — it authorizes via OAuth (no API key in config) and runs this model with the run_model tool using the endpoint bytedance/seedance-v2.5/text-to-video.

MCP client config (Claude Desktop, Cursor)

{
  "mcpServers": {
    "modelrunner": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://mcp.modelrunner.run/mcp"]
    }
  }
}

Claude Code

claude mcp add --transport http modelrunner https://mcp.modelrunner.run/mcp

Then ask your assistant, for example: “Run bytedance/seedance-v2.5/text-to-video on ModelRunner to generate video”. MCP setup guide.

Model Detayları

Seedance 2.5, yazılı bir prompt’u ona uyan bir ses bandıyla birlikte bir video klibe dönüştürür: diyalog, ortam sesi ve ses efektleri sonradan dublajlanmak yerine görüntüyle birlikte üretilir. Öne çıkan yeteneği süre: tek bir çekim 24 fps’de 30 saniyeye kadar sürebilir, yani 2.0 neslinin tavanının iki katı; böylece bütün bir an, parçalar birbirine kurgulanmadan tek çekimde akar. Çıktı 480p ya da 720p’dir ve geniş ekran, kare, dikey ve 21:9 sinematik biçimleri kapsayan yedi en-boy oranında alınabilir. `duration` alanını varsayılan `-1` değerinde bırakırsanız model, tarif ettiğiniz çekime uygun tam saniyelik bir süreyi 4 ile 30 saniye arasında kendisi seçer.

## En uygun olduğu işler - Birkaç saniyeden uzun sürmesi gereken tek çekimlik klipler: bütün bir aksiyon anı, yürürken konuşma, kısa bir sahne - Ayrı bir ses çalışması gerektirmeden sesli gelmesi gereken sosyal medya, reklam ve fragman kurguları - Sesin ağız hareketiyle örtüşmesi gereken diyalog ve şarkı çekimleri - Ortam sesi baştan gömülü gelen tanıtıcı çekimler, b-roll ve atmosfer planları - Aynı sahneyi farklı yerleşimler için 9:16, 1:1 ve 16:9 olarak kurgulamak

## Şu durumlarda başka bir model seçin - 1080p gerekiyorsa: bu nesil en fazla 720p verir, bu yüzden bunun yerine `bytedance/seedance-v2/text-to-video` kullanın - Elinizdeki bir fotoğrafı canlandırmak istiyorsanız: durağan görselinizi ilk kare olarak alan `bytedance/seedance-v2/image-to-video` kullanın - Çalıştırmaların sabit bir seed ile tekrarlanabilir olması gerekiyorsa: bu nesilde seed girdisi yok - Kilitli, tamamen sabit bir kamera gerekiyorsa: burada kamera kilidi sunulmuyor, bu yüzden çerçevelemeyi prompt’ta tarif edin

## İpuçları - Sesi de görüntü kadar açık tarif edin (çığlık atan martılar, çakıllara vuran dalgalar); ses bandı prompt’u izler - Maliyet önemliyse `duration` değerini kendiniz belirleyin: faturalandırma bitmiş videonun saniyesi üzerindendir, `-1` ise toplamı 30 saniyeye kadar açık bırakır - Denemeleri `480p` ile yapın, beğendiğiniz sonucu `720p` ile yeniden çalıştırın — maliyeti belirleyen iki etken çözünürlük ve süredir - Sessiz bir klip için `generate_audio` alanını false yapın; fiyat her iki durumda da aynıdır

## Sınırlamalar - Hızlı hareket ve kalabalıklar 720p’de 480p’ye göre daha iyi bütünlüğünü korur - Çok sayıda ayrı çekimi üst üste yığan bir prompt, uzun bir çekim boyunca karakterde, kostümde ve ışıkta sapmaya yol açabilir - Görüntünün içine basılan metinde sık sık yazım hatası olur

ModelRunner JavaScript client ile çalıştırmak için: ```js import { modelrunner } from "@modelrunner/client";

const result = await modelrunner.subscribe("bytedance/seedance-v2.5/text-to-video", { input: { prompt: "A lighthouse keeper climbs the spiral stair at dawn, lamp glass catching the first light, gulls calling outside as the storm clears", resolution: "720p", aspect_ratio: "9:16", duration: 12, generate_audio: true, }, }); ```

Seedance 2.5 Text to Video API

Model girdisi

Input

Additional Settings

Model çıktısı

Output

Örnek istekler

Örnekler

Model fiyatlandırması

Fiyatlandırma

Seedance 2.5 Text to Video API

cURL

JavaScript

Python

Input parameters

Use Seedance 2.5 Text to Video from Claude & Cursor (MCP)

MCP client config (Claude Desktop, Cursor)

Claude Code

Model Detayları

Model Detayları