ComfyCloudMiniMaxH3ImageToVideoNode - ComfyUI Built-in Node Documentation - ComfyUI

Runs the MiniMax H3 image-to-video workflow on a Comfy Cloud GPU and returns a video with a matching soundtrack. The still becomes the first frame. Feeding a clip’s closing frame back in continues the sequence, so chaining several runs builds a longer shot. This is a beta node and is billed by GPU second.

Inputs

Parameter	Description	Data Type	Required	Range
`first_frame`	The image the video starts from. It is uploaded at up to 2048x2048 pixels before the workflow runs.	IMAGE	Yes	-
`prompt`	The text description of the motion and the sounds the shot should carry. Leading and trailing whitespace is removed before submission.	STRING	Yes	1 to 4096 characters
`seed`	Random seed that controls the generated result (default: 42).	INT	Yes	0 to 18446744073709551615
`aspect_ratio`	Aspect ratio of the generated video (default: `"1:1"`).	COMBO	Yes	`"1:1"` `"3:4"` `"2:3"` `"3:2"` `"4:3"` `"16:9"` `"9:16"` `"21:9"`
`resolution`	Frame size budget (default: `"480p"`). 720p costs roughly twice the GPU-seconds of 480p.	COMBO	Yes	`"480p"` `"720p"`
`duration_seconds`	Length in seconds (default: 5). The pipeline quantises to 17-frame steps at 24 fps, so the clip lands within about two thirds of a second of this value.	INT	Yes	5 to 15

Outputs

Output Name	Description	Data Type
`VIDEO`	The generated video with its soundtrack.	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): d3657e266f2b9f789414a106ef66b6c64ec314d27c3b6a99c3a6a9b8584abf61