kling/kling-motion-control
Kling 3.0 motion control: transfer motion from a reference video to any character image with improved consistency and quality.
Pricing
kling/
kling-motion-control
Pricing for Synexa AI models works differently from other providers. Instead of being billed by time, you are billed by input and output, making pricing more predictable.
For example, generating 100 videos should cost around $200.00.
Check out our docs for more information about how per-request pricing works on Synexa.
| Provider | Price ($) | Saving (%) |
|---|---|---|
| Synexa | $2.0000 | - |
| replicate | $3.0000 | 33.3% |
Readme
Kling 3.0 motion control
Transfer character motion from a reference video to any image with improved consistency and quality.
About
Kling 3.0 motion control generates videos where character actions match a reference video while the visual appearance is based on a reference image. Provide an image of a character and a video showing the motion you want, and the model produces a new video combining the character's look with the reference motion.
This V3.0 update brings improved element consistency and higher quality output compared to V2.6, with better preservation of character identity and smoother motion transfer.
How it works
- Upload a reference image — the character whose appearance you want in the output
- Upload a reference video — the motion you want the character to follow
- Optionally add a text prompt to guide elements, backgrounds, and motion effects
The model transfers the motion from your reference video onto the character from your image, generating smooth, natural-looking animation. Output is rendered at 720p with the character oriented to match the reference image (max 10s output duration).
Use cases
- Dance and gesture transfer: make any character perform a specific dance or gesture from a reference video
- Character animation: bring illustrated or photographed characters to life with realistic motion
- Social media content: combine a subject's look with trending motion clips
- Marketing and advertising: animate product mascots or brand characters with specific movements
Tips
- Keep character proportions consistent between the image and the reference video. Avoid a full-body reference video with a half-body image.
- Make sure the character's entire body and head are clearly visible and not obstructed in the reference image.
- Use reference videos with clear, steady movements. The model handles moderate motion well but may struggle with very fast or chaotic movement.
- Match your text prompt to the motion being transferred to help the model understand context.
Input limits
- Reference image: .jpg/.jpeg/.png, max 10MB, dimensions 340px-3850px, aspect ratio 1:2.5 to 2.5:1
- Reference video: .mp4/.mov, max 100MB, 3-30 seconds duration
- Duration and aspect ratio are inherited from the reference video and character image — not configurable
- Audio is on by default (set keep_original_sound=false for silent output)