TextEncodeZImageOmni - ComfyUI Built-in Node Documentation

TextEncodeZImageOmni encodes a text prompt together with up to three optional reference images into a conditioning format for image generation models. The prompt is tokenized and encoded with the CLIP model, and each connected image can optionally be processed by a vision encoder and/or a VAE so that visual references are embedded alongside the text. This node is marked as experimental.

Inputs

Parameter	Description	Data Type	Required	Range
`clip`	The CLIP model used for tokenizing and encoding the text prompt.	CLIP	Yes
`image_encoder`	An optional vision encoder model. If provided, it is used to encode the input images, and the resulting embeddings are added to the conditioning.	CLIP_VISION	No
`prompt`	The text prompt to be encoded. Supports multiline input and dynamic prompts.	STRING	Yes
`auto_resize_images`	When enabled (default: True), input images are automatically resized before VAE encoding so their total pixel area is close to 1024x1024, with dimensions rounded to multiples of 8.	BOOLEAN	No	True False
`vae`	An optional VAE model. If provided, it is used to encode the input images into latent representations, which are added to the conditioning as reference latents.	VAE	No
`image1`	The first optional reference image.	IMAGE	No
`image2`	The second optional reference image.	IMAGE	No
`image3`	The third optional reference image.	IMAGE	No

Note: The node accepts a maximum of three images (image1, image2, image3). The image_encoder and vae inputs are only used when at least one image is provided; when both are connected, each image is processed by both. When auto_resize_images is True and a vae is connected, images are resized to have a total pixel area close to 1024x1024 before encoding.

Outputs

Output Name	Description	Data Type
`CONDITIONING`	The final conditioning output. It contains the encoded text prompt and, when images are provided, may include encoded image embeddings, reference latents, and extra text embeddings derived from the image placeholder template.	CONDITIONING

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): b40a3150f536b6f37e2b53e6d9992fcb4fd32dceb540c0a76773a7ba1af9a7b8

​Inputs

​Outputs

Inputs

Outputs