Seedance 1.5 Video Generation — Create Task API Reference

Use this endpoint to submit a video generation task with ByteDance’s Seedance 1.5 model. Unlike simpler text-only video APIs, Seedance 1.5 accepts a content array that lets you mix text prompts and image inputs in a single request. You can generate video from text alone, from a single reference frame, from first and last frames, or from multiple reference images. Audio generation is supported on all model tiers. The API responds immediately with a task id — you then poll the Query Video Task endpoint until the video is ready.

Video generation is asynchronous. The create endpoint returns a task id with status: "queued". Use the Query Video Task endpoint to check progress and retrieve the final video_url.

Base URL & Endpoint

Base URL:  https://zcbservice.aizfw.cn/kyyReactApiServer
Endpoint:  POST /v1/seedance/videos
Full URL:  https://zcbservice.aizfw.cn/kyyReactApiServer/v1/seedance/videos

All requests must include your API key as a Bearer token:

Authorization: Bearer YOUR_API_KEY

Supported Models

Model	Resolution	Audio	Image-to-Video
`seedance_1_5_pro_480p`	480p	✅	✅
`seedance_1_5_pro_720p`	720p	✅	✅
`seedance_1_5_pro_1080p`	1080p	✅	✅

Request Parameters

string

required

The Seedance 1.5 model to use. Accepted values:

seedance_1_5_pro_480p
seedance_1_5_pro_720p
seedance_1_5_pro_1080p

array

required

An array of content objects that define the prompt and any input images. This is not a simple string — it must be an array of typed objects (see the Content Array Structure section below).At minimum, include one text object. You may also include image objects to enable image-guided generation modes.

string

Output video aspect ratio. Accepted values: 16:9, 4:3, 1:1, 3:4, 9:16, 21:9, adaptive.Defaults:

Text-to-video: 16:9 (Seedance 1.5 Pro defaults to adaptive)
Image-to-video: adaptive

integer

Video duration in seconds.

Range: 2–12 seconds (default 5)
Seedance 1.5 Pro also supports 4–12 seconds
Set to -1 to let the model automatically choose the best duration (4–12 s)

boolean

Whether to include synthesised audio in the output video. Defaults to true.Audio generation is only available on Seedance 1.5 Pro series models.

Content Array Structure

The content array supports two object types: text and image_url.

Text Content Object

string

required

Must be "text".

string

required

The video generation prompt. Maximum length: 500 Chinese characters or 1000 English words.

Example text object

{
  "type": "text",
  "text": "A kitten yawning at the camera"
}

Image Content Object

string

required

Must be "image_url".

string

required

A publicly accessible URL for the input image.

string

The role of this image in the generation. Controls which generation mode is used:

first_frame (or omitted) — use as the first frame of the video (image-to-video mode)
last_frame — use as the last frame of the video (must be paired with a first_frame image)
reference_image — use as a stylistic reference (1–4 images supported)

Image requirements:

Formats: JPEG, PNG, WebP, BMP, TIFF, GIF (Seedance 1.5 Pro also supports HEIC/HEIF)
Aspect ratio: between 0.4 and 2.5 (width ÷ height)
Dimensions: 300 px – 6000 px on each side
File size: less than 30 MB

Example image object

{
  "type": "image_url",
  "image_url": {
    "url": "https://example.com/first-frame.jpg"
  },
  "role": "first_frame"
}

Generation Scenarios

The image roles in the content array determine which generation mode is active. These scenarios are mutually exclusive — you cannot mix first-frame, last-frame, and reference image roles in the same request.

Scenario	Content array contents
Text-to-video	One text object only
Image-to-video (first frame)	One text object + one image with `role: "first_frame"`
Image-to-video (first + last frame)	One text object + one `first_frame` image + one `last_frame` image
Reference image guidance	One text object + 1–4 images with `role: "reference_image"`

The three image-guided scenarios are mutually exclusive. Do not mix first_frame, last_frame, and reference_image roles in a single request.

Response Fields

string

Unique identifier for the video generation task. Task records are saved for 1 day — save this value to query the result in time.

string

Object type. Always "video".

integer

Unix timestamp (seconds) of when the task was created.

string

The model name you specified in the request.

string

Initial task status. On successful creation this is always "queued".

Code Examples

curl --request POST \
  --url https://zcbservice.aizfw.cn/kyyReactApiServer/v1/seedance/videos \
  --header 'Authorization: Bearer YOUR_API_KEY' \
  --header 'Content-Type: application/json' \
  --data '{
    "model": "seedance_1_5_pro_720p",
    "content": [
      {
        "type": "text",
        "text": "A kitten yawning at the camera, soft natural lighting"
      }
    ],
    "ratio": "16:9",
    "duration": 5,
    "generate_audio": true
  }'

curl --request POST \
  --url https://zcbservice.aizfw.cn/kyyReactApiServer/v1/seedance/videos \
  --header 'Authorization: Bearer YOUR_API_KEY' \
  --header 'Content-Type: application/json' \
  --data '{
    "model": "seedance_1_5_pro_720p",
    "content": [
      {
        "type": "text",
        "text": "The kitten slowly opens its eyes and looks around curiously"
      },
      {
        "type": "image_url",
        "image_url": {
          "url": "https://example.com/kitten-sleeping.jpg"
        },
        "role": "first_frame"
      }
    ],
    "ratio": "16:9",
    "duration": 5,
    "generate_audio": true
  }'

import requests

url = "https://zcbservice.aizfw.cn/kyyReactApiServer/v1/seedance/videos"

headers = {
    "Authorization": "Bearer YOUR_API_KEY",
    "Content-Type": "application/json",
}

# Text-to-video example
payload = {
    "model": "seedance_1_5_pro_720p",
    "content": [
        {
            "type": "text",
            "text": "A kitten yawning at the camera, soft natural lighting",
        }
    ],
    "ratio": "16:9",
    "duration": 5,
    "generate_audio": True,
}

response = requests.post(url, headers=headers, json=payload)
data = response.json()

task_id = data["id"]
print(f"Task created: {task_id} — status: {data['status']}")

Example success response:

{
  "id": "video_abc123def456",
  "object": "video",
  "created": 1761635478,
  "model": "seedance_1_5_pro_720p",
  "status": "queued"
}

Tips

Use the 480p model (seedance_1_5_pro_480p) for fast previews before committing to a 720p or 1080p render. 480p tasks typically complete in 30 seconds to 2 minutes.

Task IDs and generated videos are retained for 1 day. Query the result and download the video promptly — both will be permanently deleted after 24 hours.

Set duration: -1 to let Seedance 1.5 Pro automatically choose the most appropriate video length (4–12 seconds) based on your prompt and input images.

Overview

Text APIs

Image APIs

Video APIs

Asset Management

Seedance 1.5 Video Generation — Create Task API Reference

Base URL & Endpoint

Supported Models

Request Parameters

Content Array Structure

Text Content Object

Image Content Object

Generation Scenarios

Response Fields

Code Examples

Tips

​Base URL & Endpoint

​Supported Models

​Request Parameters

​Content Array Structure

​Text Content Object

​Image Content Object

​Generation Scenarios

​Response Fields

​Code Examples

​Tips

Base URL & Endpoint

Supported Models

Request Parameters

Content Array Structure

Text Content Object

Image Content Object

Generation Scenarios

Response Fields

Code Examples

Tips