ImageOnlyCheckpointSave - ComfyUI Built-in Node Documentation

This node saves a checkpoint file that bundles a model together with its CLIP vision encoder and VAE.

This node saves a checkpoint file that bundles a model together with its CLIP vision encoder and VAE. The file is written in safetensors format to the output directory, using the given filename prefix, so the image-related components of a model can be stored as a single checkpoint.

Inputs

Parameter	Description	Data Type	Required	Range
model	The model to be saved in the checkpoint	MODEL	Yes	-
clip_vision	The CLIP vision encoder to be saved in the checkpoint	CLIP_VISION	Yes	-
vae	The VAE (Variational Autoencoder) to be saved in the checkpoint	VAE	Yes	-
filename_prefix	The prefix for the output filename (default: “checkpoints/ComfyUI”)	STRING	Yes	-
prompt	Hidden parameter that receives the workflow prompt data	PROMPT	No	-
extra_pnginfo	Hidden parameter that receives additional PNG metadata	EXTRA_PNGINFO	No	-

Outputs

Output Name	Description	Data Type
-	This node does not return any outputs	-

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): 8ff4b3a78d8da523eaa5f784f847e954ba73b4d6037e748dcce592b447fcdee9