Nodes/ComfyUl-UCloud/Modelverse Kling V3
ComfyUI Node

Modelverse Kling V3

A ComfyUI node in UCLOUD_MODELVERSE/Kling with 16 inputs and 2 outputs.

By ucloud·Created about a year ago·Updated 2 months ago· 3
Modelverse Kling V3
  • client
  • first_frame_image
  • last_frame_image
  • url
  • task_id
prompt
negative_prompt
kling_v3_typeauto
first_frame_url
last_frame_url
reference_video_url
aspect_ratio16:9
duration5
modestd
soundoff
shot_type
character_orientationimage
keep_original_soundno
CategoryUCLOUD_MODELVERSE/Kling

Inputs (16)

NameTypeDefaultDescription
clientMODELVERSE_API_CLIENT
promptSTRINGText prompt describing the desired video
negative_promptoptSTRINGNegative prompt to avoid unwanted content
kling_v3_typeoptCOMBOautoTask type; auto infers from inputs (image→i2v, video→motion_control)
first_frame_imageoptIMAGEFirst frame image (image-to-video / motion control)
first_frame_urloptSTRINGFirst frame URL (use either this OR first_frame_image, not both)
last_frame_imageoptIMAGELast frame image (image-to-video)
last_frame_urloptSTRINGLast frame URL (use either this OR last_frame_image, not both)
reference_video_urloptSTRINGReference video URL (motion control; auto-selects motion_control when set)
aspect_ratiooptCOMBO16:9Output aspect ratio
durationoptINT53–15Video duration in seconds (3–15)
modeoptCOMBOstdstd: 720p, pro: 1080p
soundoptCOMBOoffGenerate synchronized audio
shot_typeoptSTRINGOptional shot type, e.g. 'multi' for multi-shot
character_orientationoptCOMBOimageMotion control: follow image or video orientation
keep_original_soundoptCOMBOnoMotion control: keep original reference video sound

Outputs (2)

NameTypeDescription
urlSTRING
task_idSTRING