Skip to main content
The Kling Motion Control node generates a video by applying the motion, expressions, and camera movements from a reference video to a character defined by a reference image and a text prompt. It allows you to control whether the character’s final orientation comes from the reference video or the reference image.

Inputs

ParameterDescriptionData TypeRequiredRange
promptA text description of the desired video. Maximum length is 2500 characters.STRINGYesN/A
reference_imageAn image of the character to animate. Minimum dimensions are 340x340 pixels. The aspect ratio must be between 1:2.5 and 2.5:1.IMAGEYesN/A
reference_videoA motion reference video used to drive the character’s movement and expression. Minimum dimensions are 340x340 pixels, maximum dimensions are 3850x3850 pixels. Duration limits depend on the character_orientation setting.VIDEOYesN/A
keep_original_soundDetermines if the original audio from the reference video is kept in the output. Default is True.BOOLEANNoN/A
character_orientationControls where the character’s facing/orientation comes from. "video": movements, expressions, camera moves, and orientation follow the motion reference video (other details via prompt). "image": movements and expressions still follow the motion reference video, but the character orientation matches the reference image (camera/other details via prompt).COMBONo"video"
"image"
modeThe generation mode to use.COMBONo"pro"
"std"
modelThe Kling model version to use. Default is "kling-v2-6".COMBONo"kling-v3"
"kling-v2-6"
Constraints:
  • The reference_video duration must be between 3 and 30 seconds when character_orientation is set to "video".
  • The reference_video duration must be between 3 and 10 seconds when character_orientation is set to "image".

Outputs

Output NameDescriptionData Type
outputThe generated video with the character performing the motion from the reference video.VIDEO
This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub

Source fingerprint (SHA-256): 3e350ca448a927367edfc90c5abb572be7ffa589c9e8e8734066af280233bd60