Kling O3 Image Edit

curl --request POST \
  --url https://api.vidgo.ai/api/generate/submit \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "kwaivgi/kling-image-o3/edit",
  "input": {
    "image_urls": [
      "https://example.com/image.png"
    ],
    "prompt": "A quiet forest under soft morning light, with natural detail.",
    "resolution": "1K",
    "size": "16:9",
    "output_format": "png",
    "n": 1
  }
}
'

{
  "code": 200,
  "data": {
    "task_id": "task-unified-example",
    "status": "running",
    "created_time": "2026-09-10T08:00:00"
  }
}

{
  "detail": "<string>"
}

{
  "detail": "<string>"
}

{
  "detail": "<string>"
}

{
  "detail": "<string>"
}

Request timeout

{
  "detail": "<string>"
}

{
  "detail": "<string>"
}

{
  "detail": "<string>"
}

{
  "detail": "<string>"
}

{
  "detail": "<string>"
}

Rework images with Kling O3 using multiple references and subject elements, with output up to 4K and support for visual variations.

​

Models

• kwaivgi/kling-image-o3/edit - Edit images using a prompt and image references.

​

Features

• Edit images using a prompt and image references.
• Accepts 1-10 images in image_urls.
• Choose from 1K, 2K, 4K output resolutions.
• Supply additional subject references through elements.
• Control the number of output images with n.

​

Output Options

​

Image Size

• size: Optional. Accepted values: auto, 16:9, 9:16, 1:1, 4:3, 3:4, 3:2, 2:3, 21:9. Default: 16:9.

​

Resolution

• resolution: Optional. Accepted values: 1K, 2K, 4K. Default: 1K.

​

Image Count

• n: Optional. Number of images to generate. Integer from 1 to 9. Default: 1. Fractional values, strings, and booleans are rejected.

​

Output Format

• output_format: Optional. Accepted values: jpeg, png, webp. Default: png.

​

Advanced Parameters

Defaults apply only when an optional input field is omitted. Explicit null or empty-string values are rejected. An empty elements array is accepted. Set model and optional callback_url at the request root. Place all workflow parameters, including output options, inside input. Send only the input fields supported by this workflow. Unsupported input fields are rejected before credits are deducted. Input URLs must use HTTP(S), include a hostname, and contain no embedded credentials.

​

Prompt

• prompt: Required. Describe the requested changes. Length: 1-2,000 characters.
• The value must not be blank when supplied.

​

Image URLs

• image_urls: Required. Provide publicly accessible image URLs. Supply 1-10 items. Each item must be an HTTP(S) URL.

​

Subject Elements

• elements: Optional. Provide subject-reference objects. Supply an array. Each item is an object.
• elements[].frontal_image_url: Optional. Use a publicly accessible HTTP(S) URL.
• elements[].reference_image_urls: Optional. Supply an array. Each item must be an HTTP(S) URL.

Refer to subjects in the prompt as @Element1, @Element2, etc. Element objects may include additional fields; no element count limit is specified.

​

Billing Factors

Billing depends on output resolution (resolution) and output image count (n). Text-to-image and image editing use the same rates. Reference images and subject elements do not add charges. Credits are deducted on submission and refunded if the task fails.