LTXVImgToVideo - ComfyUI Built-in Node Documentation

LTXVImgToVideo converts an input image into a video latent representation for video generation models. It resizes the image to the requested width and height, encodes it with the VAE, and places the encoded frames at the start of a video-sized latent of zeros. The strength control determines how much of the original image content is preserved versus modified during video generation.

Inputs

Parameter	Description	Data Type	Required	Range
`positive`	Positive conditioning prompts for guiding the video generation	CONDITIONING	Yes	-
`negative`	Negative conditioning prompts for avoiding certain elements in the video	CONDITIONING	Yes	-
`vae`	VAE model used for encoding the input image into latent space	VAE	Yes	-
`image`	Input image to be converted into video frames	IMAGE	Yes	-
`width`	Output video width in pixels (default: 768, step: 32)	INT	No	64 to MAX_RESOLUTION
`height`	Output video height in pixels (default: 512, step: 32)	INT	No	64 to MAX_RESOLUTION
`length`	Number of frames in the generated video (default: 97, step: 8)	INT	No	9 to MAX_RESOLUTION
`batch_size`	Number of videos to generate simultaneously (default: 1)	INT	No	1 to 4096
`strength`	Control over how much of the original image content is preserved in the first frames of the generated video. A value of 1.0 preserves the original image completely, while 0.0 allows maximum modification (default: 1.0)	FLOAT	No	0.0 to 1.0

Note: width and height change in steps of 32 pixels, and length changes in steps of 8 frames, matching the video latent compression (32x in the spatial dimensions and 8x in the temporal dimension). The video latent contains ((length - 1) // 8) + 1 frames.

Outputs

Output Name	Description	Data Type
`positive`	The positive conditioning passed through unchanged for use with the generated latent	CONDITIONING
`negative`	The negative conditioning passed through unchanged for use with the generated latent	CONDITIONING
`latent`	Video latent representation containing the encoded image frames and a noise mask that controls how strongly the conditioning is applied during video generation	LATENT

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): 4ebc7f80b4d9ac3329e3349c7048885de22b827b5bdd102976687afd7e07a16b

​Inputs

​Outputs

Inputs

Outputs