LatentCut - ComfyUI Built-in Node Documentation

The LatentCut node extracts a specific section from latent samples along a chosen dimension. It cuts out a portion of the latent representation by specifying the dimension (x, y, or t), the starting position, and how much to extract. The node supports both positive and negative indexing and automatically adjusts the extraction amount so it stays within the available bounds.

Inputs

Parameter	Description	Data Type	Required	Range
`samples`	The input latent samples to extract from	LATENT	Yes	-
`dim`	The dimension along which to cut the latent samples. “x” cuts along the last axis (typically width), “y” along the second-to-last axis (typically height), and “t” along the third-to-last axis (typically frames in video latents)	COMBO	Yes	”x" "y" "t”
`index`	The starting position for the cut (default: 0). Positive values count from the start, negative values count from the end. The node clamps the index to stay within the valid range of the latent samples	INT	Yes	-16384 to 16384
`amount`	The number of elements to extract along the specified dimension (default: 1). Must be at least 1. The node automatically reduces this value if it would exceed the available data beyond the starting index	INT	Yes	1 to 16384

Note: The index and amount values are adjusted to fit the actual size of the latent along the selected dimension. If index is larger than the dimension size, it is clamped to the last valid position. If index is negative, it is clamped to the dimension size in absolute terms, and amount is limited so it does not go past the end of the data.

Outputs

Output Name	Description	Data Type
`output`	The extracted portion of the latent samples	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): 7682de2644a4b85dba0571406f9f9802eca7caab09dc2ccf1ac91dc13b41bcdf

​Inputs

​Outputs

Inputs

Outputs