Kling 2.6
Kling v2.6 Motion Control Video Generation
- Kling motion control model (reference image + reference video)
- Called via unified endpoint
/v1/videos/generations - Supports image / video character orientation, max duration 10s / 30s respectively
- Asynchronous task — returns task_id on submission
POST
Documentation Index
Fetch the complete documentation index at: https://docs.apimart.ai/llms.txt
Use this file to discover all available pages before exploring further.
Authentication
All requests require Bearer Token authenticationGet your API Key:Visit the API Key Management page to obtain your API KeyAdd the following header to each request:
Request Parameters
Model name:
kling-v3-motion-control or kling-v2-6-motion-controlText prompt describing the desired motion, camera movement, and styleOptional but recommended — more specific descriptions produce more stable resultsExample:
"The character dances following the reference video, smooth motion, realistic style"Reference image URLMust be a publicly accessible link
Reference video URLMust be a publicly accessible direct link; mp4/mov recommended, under 100MB
Whether to retain the original audio track from the reference videoOptions:
yes: Keep original audio (default)no: Do not keep original audio
Character orientation controlOptions:
image: Use the character orientation from the reference image (reference video duration:3~10s)video: Use the character orientation from the reference video (reference video duration:3~30s)
Generation modeOptions:
std: Standard mode (balanced speed and quality)pro: High-quality mode (higher latency)
Watermark control object (optional)
Duration Rules
| Condition | Allowed Reference Video Duration |
|---|---|
character_orientation = image | 3s ~ 10s |
character_orientation = video | 3s ~ 30s |
Billing duration is determined by the actual duration probed from
video_url by the server, not a client-side estimate.Response
Response status code, 200 on success
Response data array