Kling Multi-Elements Video Editing

Edit videos using natural language, supporting adding, replacing, and removing elements in videos.

Workflow

Multi-elements video editing requires the following steps:

1. Initialize Video - Upload the video to be edited
2. Mark Selection (Optional) - Mark the video elements to be edited
3. Preview Selection (Optional) - Preview the marked selection effect
4. Create Task - Submit the video editing task

---

1. Initialize Video

Endpoint: POST /kling/v1/videos/multi-elements/init-selection

Description: Initialize the video to be edited and get a session ID for subsequent operations.

Authentication: Bearer Token

Authorization: Bearer YOUR_API_TOKEN

Request Parameters (Body)

Field	Type	Required	Default	Description
video_id	string	Optional	empty	Video ID from history, only supports videos generated within 30 days. Choose either this or video_url.
video_url	string	Optional	-	Video URL. Only supports MP4 and MOV formats. Duration must be ≥2s and ≤5s, or ≥7s and ≤10s. Width/height must be between 720px and 2160px. Only supports 24, 30, or 60fps. Choose either this or video_id.

Response Parameters

Field	Type	Description
code	integer	Error code (0 indicates success)
message	string	Error message
request_id	string	Request ID
data.status	integer	Rejection code, non-zero means recognition failed
data.session_id	string	Session ID, valid for 24 hours
data.fps	number	Parsed video frame rate
data.original_duration	integer	Parsed video duration
data.width	integer	Parsed video width
data.height	integer	Parsed video height
data.total_frame	integer	Parsed video total frames
data.normalized_video	string	Normalized video URL

Response Example

{
  "code": 0,
  "message": "success",
  "request_id": "req_123456",
  "data": {
    "status": 0,
    "session_id": "session_abc123",
    "fps": 30.0,
    "original_duration": 5000,
    "width": 1280,
    "height": 720,
    "total_frame": 150,
    "normalized_video": "https://example.com/normalized_video.mp4"
  }
}

---

2. Add Selection

Endpoint: POST /kling/v1/videos/multi-elements/add-selection

Description: Mark the area to be edited in video frames.

Request Parameters (Body)

Field	Type	Required	Default	Description
session_id	string	Required	-	Session ID
frame_index	integer	Required	-	Frame number, max 10 marked frames supported
points	array	Required	-	Point coordinate array
points[].x	number	Required	-	X coordinate, range [0, 1]
points[].y	number	Required	-	Y coordinate, range [0, 1], [0,0] represents top-left corner

Response Parameters

Field	Type	Description
data.status	integer	Rejection code
data.session_id	string	Session ID
data.res.frame_index	integer	Frame number
data.res.rle_mask_list	array	RLE mask list containing image segmentation results

---

3. Delete Selection

Endpoint: POST /kling/v1/videos/multi-elements/delete-selection

Description: Delete marked selection points.

Request Parameters (Body)

Field	Type	Required	Description
session_id	string	Required	Session ID
frame_index	integer	Required	Frame number
points	array	Required	Point coordinates to delete, must match exactly with add selection

---

4. Clear Selection

Endpoint: POST /kling/v1/videos/multi-elements/clear-selection

Description: Clear all marked selections.

Request Parameters (Body)

Field	Type	Required	Description
session_id	string	Required	Session ID

---

5. Preview Selection

Endpoint: POST /kling/v1/videos/multi-elements/preview-selection

Description: Preview the video effect with marked selections.

Request Parameters (Body)

Field	Type	Required	Description
session_id	string	Required	Session ID

Response Parameters

Field	Type	Description
data.res.video	string	Video URL with mask
data.res.video_cover	string	Video cover URL with mask
data.res.tracking_output	string	Mask result for each frame

---

6. Create Task

Endpoint: POST /kling/v1/videos/multi-elements

Description: Create a multi-elements video editing task.

Request Parameters (Body)

Field	Type	Required	Default	Description
model_name	string	Optional	kling-v1-6	Model name
session_id	string	Required	-	Session ID
edit_mode	string	Required	-	Edit mode: addition, swap, removal
image_list	array	Optional	empty	Reference image list. Required for addition (1-2 images), required for swap (1 image), not needed for removal.
image_list[].image	string	Required	-	Image Base64 or URL
prompt	string	Required	-	Positive text prompt, max 2500 characters. Use <<<video_1>>> for video, <<<image_1>>> for image.
negative_prompt	string	Optional	empty	Negative text prompt, max 2500 characters
mode	string	Optional	std	Generation mode: std, pro
duration	string	Optional	5	Video duration: 5 or 10 seconds
callback_url	string	Optional	empty	Callback URL
external_task_id	string	Optional	empty	Custom task ID

Recommended Prompt Templates

Addition:

Using the context of <<<video_1>>>, seamlessly add [subject] from <<<image_1>>> to [location]

Swap:

Swap [new element] from <<<image_1>>> for [original element] from <<<video_1>>>

Removal:

Delete [element description] from <<<video_1>>>

Response Parameters

Field	Type	Description
data.task_id	string	Task ID
data.task_status	string	Task status
data.session_id	string	Session ID
data.created_at	integer	Task creation time (ms)
data.updated_at	integer	Task update time (ms)

---

7. Query Task

Endpoint: GET /kling/v1/videos/multi-elements/:task_id

Description: Query the status and result of a multi-elements video editing task.

Response Parameters

Field	Type	Description
data.task_id	string	Task ID
data.task_status	string	Task status
data.task_status_msg	string	Task status message
data.task_result.videos	array	Generated video list
data.task_result.videos[].id	string	Video ID
data.task_result.videos[].session_id	string	Session ID
data.task_result.videos[].url	string	Video URL
data.task_result.videos[].duration	string	Video duration (seconds)