MiniMax H3 Reference to Video

Name: MiniMax H3 Reference to Video
Author: Nodespell Team

Official

Generate a 5–15 second video guided by mixed image, video, and audio references. Choose MiniMax H3 Reference to Video when recurring subjects, motion

Nodespell AI

AI / Video / Minimax

Generate a 5–15 second video guided by mixed image, video, and audio references. Choose MiniMax H3 Reference to Video when recurring subjects, motion examples, or sound references need distinct roles in the same shot. Up to 12 reference files can be combined, with adaptive framing and output up to 4K.

Inputs (4)

Reference Images

Object

Subject or style images addressed in the prompt as Image 1, Image 2, and so on.

Multi InputMin: 0Max: 100

Reference Videos

Object

Motion or scene videos addressed in the prompt as Video 1, Video 2, and so on.

Multi InputMin: 0Max: 100

Reference Audio

Object

Audio clips addressed in the prompt as Audio 1, Audio 2, and so on. Audio cannot be the only reference type.

Multi InputMin: 0Max: 100

Prompt

String

Text prompt describing the scene, motion, camera direction, and audio intent.

RequiredMulti InputMin: 0Max: 100

Parameters (4)

Prompt

String

Text prompt describing the scene, motion, camera direction, and audio intent.

Required

Default:

Duration

Number

Generated video duration from 5 to 15 seconds.

Default: 5

Resolution

String

Output resolution. Higher resolutions preserve more detail and take more credits.

Default: 2K

Aspect Ratio

String

Output framing, or adaptive to follow the supplied references.

Default: adaptive

Outputs (1)

Output

Inferred

Generated video output.

Nodespell Team

Creator profile

Type

Node

Status

Official

Package

Nodespell AI

Keywords (8)

Video GenerationPrompt ConditioningAspect ControlResolution ControlLength ControlReference To Video

Use in Workflow