Kling 3.0 & 3.0 Omni

curl --request POST \
  --url https://{api_domain}/v3/kling/text-to-video/kling-3.0

const options = {method: 'POST'};

fetch('https://{api_domain}/v3/kling/text-to-video/kling-3.0', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

<?php

$curl = curl_init();

curl_setopt_array($curl, [
  CURLOPT_URL => "https://{api_domain}/v3/kling/text-to-video/kling-3.0",
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_ENCODING => "",
  CURLOPT_MAXREDIRS => 10,
  CURLOPT_TIMEOUT => 30,
  CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
  CURLOPT_CUSTOMREQUEST => "POST",
]);

$response = curl_exec($curl);
$err = curl_error($curl);

curl_close($curl);

if ($err) {
  echo "cURL Error #:" . $err;
} else {
  echo $response;
}

package main

import (
	"fmt"
	"net/http"
	"io"
)

func main() {

	url := "https://{api_domain}/v3/kling/text-to-video/kling-3.0"

	req, _ := http.NewRequest("POST", url, nil)

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}

require 'uri'
require 'net/http'

url = URI("https://{api_domain}/v3/kling/text-to-video/kling-3.0")

http = Net::HTTP.new(url.host, url.port)
http.use_ssl = true

request = Net::HTTP::Post.new(url)

response = http.request(request)
puts response.read_body

POST

kling

text-to-video

kling-3.0

Kling 3.0 & 3.0 Omni

curl --request POST \
  --url https://{api_domain}/v3/kling/text-to-video/kling-3.0

const options = {method: 'POST'};

fetch('https://{api_domain}/v3/kling/text-to-video/kling-3.0', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

<?php

$curl = curl_init();

curl_setopt_array($curl, [
  CURLOPT_URL => "https://{api_domain}/v3/kling/text-to-video/kling-3.0",
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_ENCODING => "",
  CURLOPT_MAXREDIRS => 10,
  CURLOPT_TIMEOUT => 30,
  CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
  CURLOPT_CUSTOMREQUEST => "POST",
]);

$response = curl_exec($curl);
$err = curl_error($curl);

curl_close($curl);

if ($err) {
  echo "cURL Error #:" . $err;
} else {
  echo $response;
}

package main

import (
	"fmt"
	"net/http"
	"io"
)

func main() {

	url := "https://{api_domain}/v3/kling/text-to-video/kling-3.0"

	req, _ := http.NewRequest("POST", url, nil)

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}

require 'uri'
require 'net/http'

url = URI("https://{api_domain}/v3/kling/text-to-video/kling-3.0")

http = Net::HTTP.new(url.host, url.port)
http.use_ssl = true

request = Net::HTTP::Post.new(url)

response = http.request(request)
puts response.read_body

Parameters and usage are aligned with the official documentation. For detailed parameters, refer directly to the official documents.

1. Overview

This document explains how to use text-to-video, image-to-video, Omni video, and motion control through the native Kling 3.0 protocol. All four capabilities use asynchronous task mode:

Call the corresponding creation endpoint.
Save the data.id from the creation response.
Use the same API Key from task creation to query task status.
After the task succeeds, retrieve and save the generated video from outputs.

1.1 API Endpoints

Capability	Create Task	Query Task
Text-to-video	`POST /v3/kling/text-to-video/kling-3.0`	`GET /v3/kling/tasks`
Image-to-video	`POST /v3/kling/image-to-video/kling-3.0`	`GET /v3/kling/tasks`
Omni video	`POST /v3/kling/omni-video/kling-3.0-omni`	`GET /v3/kling/tasks`
Motion control	`POST /v3/kling/motion-control/kling-3.0`	`GET /v3/kling/tasks`

The full addresses in the examples in this document are https://{api_domain} plus the endpoint paths in the table, where {api_domain} is the API domain you use.

1.2 Authentication

All requests must include an API Key. POST requests must also specify the JSON content type:

Authorization: Bearer <YOUR_API_KEY>
Content-Type: application/json

1.3 Response Structure

data in the creation response is an object with the following common fields:

Field	Type	Description
`id`	string	Task ID generated by the Kling system, used for subsequent queries
`external_id`	string	Echo of `options.external_task_id`; may be omitted if not set
`status`	string	Task status after creation, usually `submitted`

data in the query response is an array, and each task contains the following fields:

Field	Type	Description
`id`	string	Task ID generated by the system
`external_id`	string	Custom task ID set by the caller, if any
`status`	string	`submitted`, `processing`, `succeeded`, or `failed`
`message`	string	Task status message; includes failure reason on failure, if any
`create_time`	integer	Creation time as Unix millisecond timestamp, if available
`update_time`	integer	Update time as Unix millisecond timestamp, if available
`outputs`	array	Generation results; a successful task usually contains one result with `type=video`

The common fields of the video result in outputs are type, id, url, watermark_url, and duration. url is the non-watermarked result URL, watermark_url is the watermarked result URL, and duration is the video duration in seconds.

2. Text-to-Video

Official parameter documentation: Text to Video

2.1 Create Task

POST https://{api_domain}/v3/kling/text-to-video/kling-3.0

Common request fields:

Field	Type	Required	Description
`prompt`	string	Yes	Video generation prompt; cannot be empty; maximum 3072 characters
`settings.multi_shot`	boolean	No	Whether to generate a multi-shot video; default `true`
`settings.resolution`	string	No	`720p`, `1080p`, or `4k`; default `720p`
`settings.aspect_ratio`	string	No	`16:9`, `9:16`, or `1:1`; default `16:9`
`settings.audio`	string	No	`off` or `native`; default `off`
`settings.duration`	integer	No	3–15 seconds; default 5 seconds
`options.callback_url`	string	No	Address that receives callback notifications when task status changes
`options.external_task_id`	string	No	Custom task ID
`options.watermark_info.enabled`	boolean	No	Whether to generate a watermarked result; default `false`

When multi_shot=true, you can describe multiple shots in the prompt using the official Kling format; the sum of the durations of each shot should equal the total video duration.

2.2 Creation Example

curl -X POST 'https://{api_domain}/v3/kling/text-to-video/kling-3.0' \
  -H 'Authorization: Bearer <YOUR_API_KEY>' \
  -H 'Content-Type: application/json' \
  -d '{
    "prompt": "A cinematic tracking shot of a red sports car driving along a coastal road at sunset",
    "settings": {
      "resolution": "1080p",
      "duration": 5,
      "audio": "native"
    },
    "options": {
      "external_task_id": "text-video-20260826-0001"
    }
  }'

After successful creation, the task ID is returned:

{
  "code": 0,
  "message": "success",
  "request_id": "provider-request-id",
  "data": {
    "id": "kling-text-task-id",
    "external_id": "text-video-20260826-0001",
    "status": "submitted",
    "create_time": 1781080778802,
    "update_time": 1781080794151
  }
}

2.3 Query Task

GET https://{api_domain}/v3/kling/tasks?task_ids={task_id}

Save data.id. If options.external_task_id is not set, data.external_id may not be returned. Use the same API Key used to create the task. If the task ID does not exist or cannot be accessed, task not found will be returned. You can also use external_task_ids to query custom task IDs:

GET https://{api_domain}/v3/kling/tasks?external_task_ids={external_task_id}

Only one of task_ids and external_task_ids can be used per request. Both support 1–50 IDs separated by English commas. data in the response is an array and preserves the order of the requested IDs.

2.4 Query Example

curl -X GET 'https://{api_domain}/v3/kling/tasks?task_ids=kling-text-task-id' \
  -H 'Authorization: Bearer <YOUR_API_KEY>'

When the task succeeds, the response is:

{
  "code": 0,
  "message": "success",
  "request_id": "provider-request-id",
  "data": [
    {
      "id": "kling-text-task-id",
      "status": "succeeded",
      "outputs": [
        {
          "type": "video",
          "url": "https://your-result-cdn.example.com/text-video.mp4",
          "duration": 5
        }
      ]
    }
  ]
}

status=processing means the task is still running; it is recommended to query every 10–30 seconds. status=failed means the task failed; do not continue polling. After the task succeeds, save the file corresponding to outputs[type=video].url promptly because the download URL may expire.

2.5 Billing

Task cost = per-second unit price for the corresponding resolution and audio tier × settings.duration

The price tier is determined by settings.resolution and settings.audio. For example, the cost for 1080p, audio=native, duration=5 is “per-second unit price of the 1080p native audio tier × 5 seconds”. If duration is not provided, the cost is billed at the default 5 seconds. The fee is deducted based on the requested duration when the task is created; if task creation fails or generation ultimately fails, the fee is automatically refunded. For the specific unit price, refer to the price page of the service you use.

3. Image-to-Video

Official parameter documentation: Image to Video

3.1 Create Task

POST https://{api_domain}/v3/kling/image-to-video/kling-3.0

Common request fields:

Field	Type	Required	Description
`contents`	array	Yes	Multimodal content list; at least one item
`contents[].type`	string	Yes	`prompt`, `first_frame`, `last_frame`, or `element`
`settings.resolution`	string	No	`720p`, `1080p`, or `4k`; default `720p`
`settings.multi_shot`	boolean	No	Whether to generate a multi-shot video; default `true`
`settings.audio`	string	No	`off` or `native`; default `off`
`settings.duration`	integer	No	3–15 seconds; default 5 seconds
`options.callback_url`	string	No	Address that receives callback notifications when task status changes
`options.external_task_id`	string	No	Custom task ID
`options.watermark_info.enabled`	boolean	No	Whether to generate a watermarked result; default `false`

contents sub-item fields:

type	Required fields	Description
`prompt`	`text`	Text prompt; maximum 3072 characters
`first_frame`	`url`	First-frame image; supports URL or Base64; required
`last_frame`	`url`	Last-frame image; supports URL or Base64; not supported alone; a first frame must also be provided
`element`	`element_id`, `id`	Subject ID and the index ID used to reference the subject in the prompt; `id` must not be duplicated within the same task; up to 3 subjects can be specified

Images support .jpg, .jpeg, and .png; file size must not exceed 50 MB; width and height must be at least 300 pixels; aspect ratio must be between 1:2.5 and 2.5:1.

3.2 Creation Example

curl -X POST 'https://{api_domain}/v3/kling/image-to-video/kling-3.0' \
  -H 'Authorization: Bearer <YOUR_API_KEY>' \
  -H 'Content-Type: application/json' \
  -d '{
    "contents": [
      {
        "type": "prompt",
        "text": "The subject turns toward the camera while the wind moves naturally through the scene"
      },
      {
        "type": "first_frame",
        "url": "https://your-cdn.example.com/first-frame.png"
      }
    ],
    "settings": {
      "resolution": "1080p",
      "duration": 5,
      "audio": "off"
    },
    "options": {
      "external_task_id": "image-video-20260826-0001"
    }
  }'

On successful creation, data.id is the task ID used for subsequent queries:

{
  "code": 0,
  "message": "success",
  "request_id": "provider-request-id",
  "data": {
    "id": "kling-image-task-id",
    "external_id": "image-video-20260826-0001",
    "status": "submitted",
    "create_time": 1781080778802,
    "update_time": 1781080794151
  }
}

3.3 Query Task

Use data.id from the creation response to query the task:

GET https://{api_domain}/v3/kling/tasks?task_ids={task_id}

When querying with task_ids, use the same API Key used to create the task. You can also query with external_task_ids using the options.external_task_id passed at creation. The query parameter supports 1–50 IDs separated by English commas.

3.4 Query Example

curl -X GET 'https://{api_domain}/v3/kling/tasks?task_ids=kling-image-task-id' \
  -H 'Authorization: Bearer <YOUR_API_KEY>'

When the task succeeds, the video URL is in outputs[type=video].url:

{
  "code": 0,
  "message": "success",
  "request_id": "provider-request-id",
  "data": [
    {
      "id": "kling-image-task-id",
      "status": "succeeded",
      "outputs": [
        {
          "type": "video",
          "url": "https://your-result-cdn.example.com/image-video.mp4",
          "duration": 5
        }
      ]
    }
  ]
}

When status=processing, continue querying every 10–30 seconds; when status=failed, stop polling and correct the input based on the error message.

3.5 Billing

Task cost = per-second unit price for the corresponding resolution and audio tier × settings.duration

The price tier is determined by settings.resolution and settings.audio; if duration is not provided, the cost is billed at the default 5 seconds. The fee is deducted based on the requested duration when the task is created; if task creation fails or generation ultimately fails, the fee is automatically refunded. For the specific unit price, refer to the price page of the service you use.

4. Omni Video

Official parameter documentation: Omni Video

4.1 Create Task

POST https://{api_domain}/v3/kling/omni-video/kling-3.0-omni

Common request fields:

Field	Type	Required	Description
`contents`	array	Yes	Multimodal content list; at least one item
`contents[].type`	string	Yes	`prompt`, `first_frame`, `last_frame`, `refer_image`, `feature_video`, `base_video`, or `element`
`settings.resolution`	string	No	`720p`, `1080p`, or `4k`; default `720p`
`settings.multi_shot`	boolean	No	Whether to generate a multi-shot video; default `true`
`settings.aspect_ratio`	string	No	`16:9`, `9:16`, or `1:1`; default `16:9`; required when there is no first-frame image or no reference video
`settings.audio`	string	No	`off`, `native`, or `original`; default `off`
`settings.duration`	integer	No	3–15 seconds; default 5 seconds
`options.callback_url`	string	No	Address that receives callback notifications when task status changes
`options.external_task_id`	string	No	Custom task ID
`options.watermark_info.enabled`	boolean	No	Whether to generate a watermarked result; default `false`

contents sub-item fields:

type	Required fields	Description
`prompt`	`text`	Text prompt; maximum 3072 characters
`first_frame`, `last_frame`, `refer_image`	`url`; `id` optional	Image assets; supports URL or Base64; `id` is used to reference the asset in the prompt and must not be duplicated within the same task
`feature_video`, `base_video`	`url`; `id` optional	Video assets; supports URL; `id` is used to reference the asset in the prompt and must not be duplicated within the same task
`element`	`element_id`, `id`	Subject ID and the index ID used to reference the subject in the prompt; `id` must not be duplicated within the same task

Images support .jpg, .jpeg, and .png; file size must not exceed 50 MB; width and height must be at least 300 pixels; aspect ratio must be between 1:2.5 and 2.5:1. Videos support .mp4 and .mov; file size must not exceed 200 MB; duration must be 3–15.5 seconds. Audio combination rules:

feature_video (feature reference video): audio can only be off (original video audio is not supported), and multi_shot can only be true.
base_video (video to edit): audio cannot be native (can be off or original), and multi-shot is not supported. audio=original is only available for requests that include a reference video.
When no video input is included, audio=off or audio=native can be used.

4.2 Creation Example

The following example uses a reference image to generate a video:

curl -X POST 'https://{api_domain}/v3/kling/omni-video/kling-3.0-omni' \
  -H 'Authorization: Bearer <YOUR_API_KEY>' \
  -H 'Content-Type: application/json' \
  -d '{
    "contents": [
      {
        "type": "prompt",
        "text": "Keep the character consistent and create a cinematic walking shot"
      },
      {
        "type": "refer_image",
        "url": "https://your-cdn.example.com/character.png"
      }
    ],
    "settings": {
      "resolution": "1080p",
      "duration": 8,
      "audio": "native

Last modified on October 15, 2026

API Basics

API Key Management

Large Language Models

Images

Video

Audio

Kling 3.0 & 3.0 Omni

1. Overview

1.1 API Endpoints

1.2 Authentication

1.3 Response Structure

2. Text-to-Video

2.1 Create Task

2.2 Creation Example

2.3 Query Task

2.4 Query Example

2.5 Billing

3. Image-to-Video

3.1 Create Task

3.2 Creation Example

3.3 Query Task

3.4 Query Example

3.5 Billing

4. Omni Video

4.1 Create Task

4.2 Creation Example

​1. Overview

​1.1 API Endpoints

​1.2 Authentication

​1.3 Response Structure

​2. Text-to-Video

​2.1 Create Task

​2.2 Creation Example

​2.3 Query Task

​2.4 Query Example

​2.5 Billing

​3. Image-to-Video

​3.1 Create Task

​3.2 Creation Example

​3.3 Query Task

​3.4 Query Example

​3.5 Billing

​4. Omni Video

​4.1 Create Task

​4.2 Creation Example

1. Overview

1.1 API Endpoints

1.2 Authentication

1.3 Response Structure

2. Text-to-Video

2.1 Create Task

2.2 Creation Example

2.3 Query Task

2.4 Query Example

2.5 Billing

3. Image-to-Video

3.1 Create Task

3.2 Creation Example

3.3 Query Task

3.4 Query Example

3.5 Billing

4. Omni Video

4.1 Create Task

4.2 Creation Example