Create video (Seedance)

This series supports OpenAI video generation formats (see the link Video generation Sora compatible formats).

Used to create video generation tasks and query status and results via task ID.

Core fields for task creation:

• model: Required. Model name.
• prompt: Optional; if not provided, content must include a non-empty text item.
• content: Optional, the Volcano Ark content array, mixing text, image, video, and audio materials.
• function_mode: Optional. Commonly used for advanced modes such as Omni-Reference.

Compatible input parameters:

The following fields are mainly used for compatibility with legacy clients and older protocols. For new integrations, please use the standard format in the “Request Example” on this page whenever possible: use the content array for text and reference materials, and use top-level duration, ratio, resolution, fps, and other fields for generation configuration.

• mode：t2v / i2v / i2v_first_last / reference_images / reference_material
• image_url / image_urls / input_reference
• end_image_url / last_image_url
• video_urls / audio_urls
• seconds / duration
• aspect_ratio / ratio / size
• quality / resolution
• fps
• generate_audio / watermark

Call details:

• When using mode=reference_material or function_mode=omni_reference, it is recommended to pass text and reference materials via content.
• content text items are written as { "type": "text", "text": "测试文字" }.
• content image, video, and audio material items use the image_url, video_url, and audio_url object formats respectively, for example { "type": "image_url", "image_url": { "url": "https://example.com/ref.png" }, "role": "reference_image", "name": "1" }.
• role can be used to identify the purpose of the material, and name can be referenced by name in prompts.
• ratio, duration, resolution, and fps will affect the final generation configuration.

Seedance 2.5 Mixed image, video, and audio reference example (the asset URLs are placeholders; replace them with actual accessible links when calling):

{
  "model": "sd-2-5",
  "ratio": "9:16",
  "fps": 24,
  "duration": 10,
  "content": [
    {
      "type": "text",
      "text": "使用 @图片1 @图片2 @图片3 @音频1 参考@视频1 生成视频"
    },
    {
      "type": "image_url",
      "image_url": {
        "url": "https://demo1.png"
      },
      "role": "reference_image",
      "name": "图片1"
    },
    {
      "type": "image_url",
      "image_url": {
        "url": "https://demo2.png"
      },
      "role": "reference_image",
      "name": "图片2"
    },
    {
      "type": "image_url",
      "image_url": {
        "url": "https://demo3.png"
      },
      "role": "reference_image",
      "name": "图片3"
    },
    {
      "type": "video_url",
      "video_url": {
        "url": "https://video.mp4"
      },
      "role": "reference_video",
      "name": "视频1"
    },
    {
      "type": "audio_url",
      "audio_url": {
        "url": "https://audio.mp3"
      },
      "role": "reference_audio",
      "name": "音频1"
    }
  ]
}

Gemini Omni invocation:

• gemini-omni can also be invoked through this Seedance task API.
• mode=t2v is used for text-to-video, mode=r2v for generation with reference images/reference materials, and mode=edit for video editing.
• Text can be placed in the text item of prompt or content; images can be placed in image_url / image_urls / reference_images / input_reference / content; videos can be placed in video_urls / content.
• The duration field can use duration or seconds, and will be automatically mapped to the 4 / 6 / 8 / 10 second options.

MiniMax / Hailuo Invocation:

• minimax-h3 can be invoked through this task API and supports text-to-video, image reference, and audio reference; text can be placed in the prompt or content text fields, and assets can be placed in fields such as content, image_urls, and audio_urls.
• hailuo-2.3 can be invoked through this task API. A first-frame image is required. It is recommended to pass the image_url item of role=first_frame in content.
• minimax-h3 commonly uses duration=5~15 and resolution=1440p; hailuo-2.3 commonly uses duration=6 or 10 and resolution=768p.

Key fields in query results:

• items[].status
• items[].content.video_url or items[].video_url
• items[].error
• items[].progress

Error fields: error.code, error.message.

{
  "error": {
    "code": "invalid_request_error",
    "message": "requires at least one image in `images`"
  }
}

Kling 4.0

Kling 4.0 calls: kling-4.0 (720p) and kling-4.0-1080p both support the Seedance content/duration/resolution/ratio/mode format, contents/settings/options, and the OpenAI video format; request bodies can be reused between the two unified video endpoints. The resolution must match the model name. Kling 4.0 can directly use the Seedance format fields content/duration/resolution/ratio/mode, or the OpenAI video format fields prompt/seconds/size/input_reference; neither requires additional contents/settings/options. The legacy contents/settings/options envelope is retained only for compatibility. Available mode values are t2v (text only), i2v (single first frame), i2v_first_last (first and last frames), reference_images (standard reference images), and reference_material (image or video references). When mode is omitted and no explicit tool is provided, reference_material is used by default if reference material is present; otherwise, t2v is used by default. Reference modes use the standard reference image role and do not silently convert explicitly specified first and last frames into reference images; for first and last frames, explicitly select i2v or i2v_first_last and omit aspect-ratio parameters. Reference audio is not currently supported. settings.audio is available only in the all-in-one mode for reference images/reference material; omit the audio switch for t2v, i2v, and i2v_first_last.

Vidu Q4

Vidu Q4: vidu-q4 is fixed at 720p, and vidu-q4-540p is fixed at 540p. Supports generating videos from 3 to 16 seconds. At least 1 image is required, with up to 15 reference images supported. Reference generation can include up to 3 MP3 audio clips, each 3 to 12 seconds long and up to 50 MB. Pure text-to-video, specified first and last frames, and reference videos are not supported. Multiple images are used as joint references and do not indicate first or last frame positions. Tasks can be created via POST /v1/videos, POST /api/v3/contents/generations/tasks, or the Vidu formats POST /ent/v2/img2video and POST /ent/v2/reference2video. For the unified endpoint, mode=reference_images is recommended; use mode=reference_material when including audio. Reference audio can use audio_urls. The Seedance format can also use the audio_url field of the standard content; the audio Boolean field controls audio generation and is used separately from the reference audio URL list. The resolution must match the model name.