Back to Nodes
Kling V3 Motion Control

Kling V3 Motion Control

Official

Transfer motion from a reference video onto a reference character image with controllable orientation, quality mode, and source-audio retention.

Nodespell AI
AI / Video / Kwaivgi

Transfer motion from a reference video onto a reference character image with controllable orientation, quality mode, and source-audio retention.

Model Overview

Kling V3 Motion Control uses one reference image plus one reference video and generates a new video where the image subject follows the source motion pattern.

Best At

  • Re-targeting motion to a different subject identity.
  • Preserving character appearance while borrowing movement rhythm.
  • Iterating quickly with standard and pro quality modes.

Limitations / Not Good At

  • Requires both image and video references.
  • Complex occlusions or rapid scene changes can reduce consistency.
  • Final quality still depends heavily on source material quality.

Ideal Use Cases

  • Character motion transfer for concept videos.
  • Previsualization workflows that need stable subject identity.
  • Movement-style experiments from existing footage.

Input & Output Format

  • Input: required image and video; optional prompt, mode, character_orientation, and keep_original_sound.
  • Output: generated video URI returned on response.

Performance Notes

  • pro mode generally yields higher visual fidelity than std.
  • Orientation mode changes whether pose behavior follows the image or video framing assumptions.
Inputs (3)

Prompt

String

Optional text instructions to guide motion style and scene details.

Multi InputMin: 0Max: 100

Reference Image

String

Reference image that defines the subject and visual scene elements.

RequiredMin: 0Max: 100

Reference Video

String

Reference motion video used to transfer action timing and movement.

RequiredMin: 0Max: 100
Parameters (4)

Prompt

String

Optional text instructions to guide motion style and scene details.

Default:

Mode

String

Quality mode for generation output.

Default: pro

Character Orientation

String

Controls whether generated character orientation follows the image or video reference convention.

Default: image

Keep Original Sound

Boolean

Keep audio from the reference video when possible.

Default: true
Outputs (1)

Output

Inferred

Output

Nodespell Team

Type

Node

Status

Official

Package

Nodespell AI

Category

AI / Video / Kwaivgi

Input

TextImageVideo

Output

Video

Keywords

Video GenerationMotion TransferPrompt ConditioningLength ControlResolution Control
Use in Workflow