Kling O1

Kling O1 uses an asynchronous task API for text, image, element, and video inputs. Submit a task, receive its ID, then query the task for status and outputs.

Create a task

Item	Value
Method	`POST`
URL	`https://api.tokenhub.belivestudios.com/omni-video/kling-o1`
Authentication	Header `Authorization: Bearer sk-your-belive-studios-key`
Content-Type	`application/json`

curl https://api.tokenhub.belivestudios.com/omni-video/kling-o1 \
  -H "Authorization: Bearer sk-your-belive-studios-key" \
  -H "Content-Type: application/json" \
  -d '{
  "contents": [
    {
      "type": "prompt",
      "text": "Have the subject walk naturally through a city street while the camera follows smoothly"
    },
    {
      "type": "first_frame",
      "url": "https://p2-kling.klingai.com/kcdn/cdn-kcdn112452/kling-tob-release_note/image_25.png",
      "id": "image_1"
    },
    {
      "type": "refer_image",
      "url": "https://p2-kling.klingai.com/kcdn/cdn-kcdn112452/kling-tob-release_note/image_25.png",
      "id": "image_2"
    }
  ],
  "settings": {
    "resolution": "1080p",
    "duration": 5,
    "aspect_ratio": "16:9",
    "audio": "off"
  },
  "options": {
    "external_task_id": "video-task-o1",
    "watermark_info": {
      "enabled": false
    }
  }
}'

Request parameters

Field	Type	Required	Default	Values	Description
`contents`	array	Yes	-	-	Collection of prompt, image, element, and video inputs.
`contents[].type`	string	Yes	-	`prompt`, `first_frame`, `last_frame`, `refer_image`, `feature_video`, `base_video`, `element`	Input type.
`contents[].text`	string	Required for `prompt`	-	-	Prompt text, up to 2,500 characters.
`contents[].url`	string	Required for image/video types	-	URL (Base64 for images)	Image or video content.
`contents[].id`	string	No (`element` requires it)	-	-	Per-task asset ID, referenced as `@id`; must be unique.
`contents[].element_id`	string	Required for `element`	-	-	System-generated element ID.
`settings`	object	No	-	-	Output settings.
`settings.audio`	string	No	`off`	`original`, `off`	Keep reference-video audio or disable audio.
`settings.resolution`	string	No	`720p`	`720p`, `1080p`	Output resolution.
`settings.aspect_ratio`	string	Conditional	`16:9`	`16:9`, `9:16`, `1:1`	Must be supplied when there is no first frame and no reference video.
`settings.duration`	integer	No	`5`	`3`–`10`	Duration in seconds. First-frame-only input without other references supports only 5 or 10 seconds.
`options`	object	No	-	-	Callback, custom task ID, and watermark settings.
`options.callback_url`	string	No	-	-	URL that receives task status callbacks.
`options.external_task_id`	string	No	-	-	Account-unique custom task ID, also usable for queries.
`options.watermark_info`	object	No	-	-	Watermark configuration.
`options.watermark_info.enabled`	boolean	No	`false`	`true`, `false`	Whether to generate a watermarked result. Custom watermarks are not supported.

`contents` input types

`type`	Purpose	Fields
`prompt`	Prompt input	`text`
`first_frame`	First-frame image	`url`; optional `id`
`last_frame`	Last-frame image	`url`; optional `id`
`refer_image`	Scene, style, or element reference image	`url`; optional `id`
`feature_video`	Feature reference video	`url`; optional `id`
`base_video`	Video to edit	`url`; optional `id`
`element`	Previously created element	`element_id`, `id`

Input and combination rules

Images: JPG, JPEG, or PNG; up to 50 MB; both dimensions at least 300 px; aspect ratio 1:2.5 to 2.5:1.
Videos: MP4 or MOV; up to 200 MB; 3–10 seconds; 700–2160 px; 24–60 fps; at most one reference video.
first_frame, last_frame, and refer_image use images. First-frame-only or first-plus-last-frame is supported; last-frame-only is not. Do not add other reference images when both first and last frames are used.
feature_video supports a first-frame definition only. base_video cannot define a first or last frame.
element supports multi-image elements only, not video-character elements. Elements cannot be used with first/last-frame generation.
Image and element counts together must not exceed 7 without a reference video, or 4 with a reference video.
Use @id in the prompt to reference an image, element, or video. IDs must be unique and should not contain one another or create ambiguity with prompt text.

Create response

{
  "code": 0,
  "message": "string",
  "request_id": "string",
  "data": {
    "id": "893605946402811985",
    "status": "submitted",
    "create_time": 1781080778802,
    "update_time": 1781080794151,
    "external_id": "video-task-o1"
  }
}

status can be submitted, processing, succeeded, or failed.

Field	Description
`code`	Error code; `0` indicates success.
`message`	Error or status message.
`request_id`	System-generated request ID for tracing.
`data.id`	System-generated task ID.
`data.status`	Current task status.
`data.create_time`, `data.update_time`	Creation and update times as Unix millisecond timestamps.
`data.external_id`	Custom task ID, when supplied.

Query a task

GET https://api.tokenhub.belivestudios.com/tasks

Provide exactly one of task_ids or external_task_ids; multiple IDs can be comma-separated.

Query parameter	Type	Required	Description
`task_ids`	string	One of the two	System task IDs, comma-separated for a batch.
`external_task_ids`	string	One of the two	Custom task IDs, comma-separated for a batch.

curl "https://api.tokenhub.belivestudios.com/tasks?task_ids=893605946402811985" \
  -H "Authorization: Bearer sk-your-belive-studios-key" \
  -H "Content-Type: application/json"

The response contains code, message, request_id, and data[]. Each task can include:

Field	Description
`id`, `status`, `message`	Task ID, state, and status message; failures include the reason.
`create_time`, `update_time`	Creation and update times as Unix millisecond timestamps.
`external_id`	Custom task ID, when supplied.
`outputs`	Generated outputs. URLs are generally removed after 30 days; save them promptly.
`billing`	Billing details for the task.

{
  "code": 0,
  "message": "string",
  "request_id": "string",
  "data": [
    {
      "id": "893605946402811985",
      "status": "succeeded",
      "message": "string",
      "create_time": 1781080778802,
      "update_time": 1781080794151,
      "external_id": "video-task-o1",
      "outputs": [
        {
          "type": "video",
          "id": "string",
          "url": "string",
          "watermark_url": "string",
          "duration": "5"
        }
      ],
      "billing": [
        {
          "charge_type": "cash",
          "cash_type": "balance",
          "amount": "string",
          "currency": "CNY",
          "list_price": "string"
        }
      ]
    }
  ]
}

outputs[] fields depend on type:

video: id, url, watermark_url, duration.
image: url, watermark_url, group_id (for grouped images).
audio: id, mp3_url, wav_url, mp3_duration, wav_duration.
voice: id, name, url, owned_by, status.
element: id, name, description, element_type, references, owned_by, status, tags.

Element references[] can contain type, role, and url; voice references can also include id, name, and owned_by. Each tags[] entry has id, name, and description.

`billing[]` field	Description
`charge_type`	`cash` for credit or `unit` for a resource package.
`cash_type`	`balance` or `test_balance`; present for `charge_type=cash`.
`amount`	Deducted amount: discounted price for credit, or points for packages.
`currency`	`CNY` or `USD`; present for cash charges.
`package_type`	`video`, `image`, or `audio`; present for unit charges.
`list_price`	Credit list price; present for cash charges.

Cursor query

POST https://api.tokenhub.belivestudios.com/tasks

This endpoint only queries asynchronous, non-real-time tasks. For the first request, use a time range and filters. For subsequent pages, pass the previous response's data.next_cursor as cursor; when cursor is supplied, start_time and end_time are ignored. The request body supports start_time, end_time, cursor, limit, and filters:

Field	Type	Default	Description
`start_time`	string	30 days before `end_time`	Unix millisecond timestamp; must be earlier than `end_time`.
`end_time`	string	Current time	Unix millisecond timestamp.
`cursor`	string	-	Previous `data.next_cursor`; when supplied, the time range is ignored.
`limit`	integer	`100`	Number of tasks, up to 500.
`filters`	array	-	Query filter collection.
`filters[].key`	string	-	`status` or `product_type`.
`filters[].values`	array	-	Status: `submitted`, `processing`, `succeeded`, `failed`; product type: `video`, `image`, `try_on`.

curl https://api.tokenhub.belivestudios.com/tasks \
  -H "Authorization: Bearer sk-your-belive-studios-key" \
  -H "Content-Type: application/json" \
  -d '{
  "limit": 100,
  "filters": [
    {
      "key": "status",
      "values": [
        "succeeded"
      ]
    },
    {
      "key": "product_type",
      "values": [
        "video"
      ]
    }
  ]
}'

Results are in data.result, with data.count, data.next_cursor, and data.has_more. Each result uses the same task fields as the ID query.

To fetch the next page, pass the previous response's data.next_cursor as cursor:

{
  "cursor": "data.next_cursor from the previous response",
  "limit": 100,
  "filters": [
    {
      "key": "product_type",
      "values": ["video"]
    }
  ]
}