Image-to-video

Generate Kling 4.0 video from a first frame, first and last frames, or multiple keyframes.

Combine prompts and keyframes in contents, with up to 10 images per task. This endpoint does not provide aspect_ratio; use reference images with the intended aspect ratio.

POST/kling/image-to-video/kling-4.0

Generate video from a first frame, first and last frames, or multiple keyframes, with up to 10 images per task. The reference image determines the aspect ratio; aspect_ratio is not provided.

Authorization

bearerAuth

AuthorizationBearer <token>

Use a platform API Key: Authorization: Bearer your-api-key.

In: header

Request Body

application/json

TypeScript Definitions

Use the request body type in TypeScript.

contents*array<|>

At least one image is required; prompts are optional. Up to 10 keyframe images. First, middle, and last frames do not all have to be provided together.

Items1 <= items

settings?

options?

Response Body

`application/json`

curl -X POST "https://api.agtcloud.cn/kling/image-to-video/kling-4.0" \
  -H "Authorization: Bearer your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [
      {
        "type": "prompt",
        "text": "Slowly animate the waves in @first while keeping the camera steady."
      },
      {
        "type": "first_frame",
        "id": "first",
        "url": "https://example.com/first.jpg"
      }
    ],
    "settings": {
      "duration": 5,
      "resolution": "720p"
    }
  }'

{
  "code": 0,
  "message": "SUCCEED",
  "request_id": "REQUEST_ID",
  "data": {
    "id": "123456789012345678",
    "status": "submitted",
    "message": "",
    "create_time": 1792713600000,
    "update_time": 1792713600000
  }
}

contents supports first_frame (first frame), middle_frame (middle frame), and last_frame (last frame). The prompt item type: "prompt" is optional. The request examples cover a first frame, first and last frames, and multiple keyframes.

Item	Requirement
Image source	`url` is required; URL or Base64 is supported
Format and size	JPG, JPEG, PNG; up to 30MB per image
Dimensions	Width and height at least 300px; aspect ratio from 2:5 to 5:2
Image count	Up to 10 per task
Asset reference ID	`id` is optional; supplied values must be unique within the task

To refer to an image in the prompt, assign an id, for example "id": "first" corresponds to @first. id is a reference within this request, not the returned generation task ID.

Common errors and troubleshooting

Response or error	What to check
HTTP 401, `Invalid token`	Check Bearer authentication and whether the API Key belongs to the requested site
HTTP 400, `contents cannot be empty`	Provide a nonempty `contents` array
HTTP 400, `contents must contain at least one of first_frame, middle_frame, last_frame`	Provide at least one keyframe image; a prompt alone is insufficient
HTTP 400, `must have a non-blank url`	Supply a nonempty `url` for the relevant keyframe
HTTP 400, `contents[0].type ... is invalid`	Check the asset at the indicated index and use a supported keyframe or prompt type
HTTP 400, `invalid_request` or `invalid JSON request body`	Check JSON syntax and field types; `contents` must be an array, not a string

HTTP 400 may return code/message/data or an error object. fail_to_fetch_task does not necessarily mean retrieval failed; read the asset validation cause in message. It may be a JSON string, not a parsed object. Keep the request ID, but do not share your API Key.

Authorization

Request Body

Response Body

200application/json

400application/json

401application/json

Image and reference rules

Common errors and troubleshooting

`application/json`

`application/json`

`application/json`