CLIPTextEncodeControlnet - ComfyUI Built-in Node Documentation - ComfyUI

The CLIP Text Encode (Controlnet) node encodes a text prompt with a CLIP model and adds the resulting text encoding to existing conditioning data. It stores the text embeddings as controlnet cross-attention parameters inside each conditioning entry, so the returned conditioning carries that extra controlnet information.

Inputs

Parameter	Description	Data Type	Required	Range
`clip`	The CLIP model used for text tokenization and encoding	CLIP	Yes	-
`conditioning`	Existing conditioning data to be combined with the CLIP text encoding	CONDITIONING	Yes	-
`text`	The text prompt to be processed by the CLIP model. Supports multiline text and dynamic prompts	STRING	Yes	-

Note: All three inputs (clip, conditioning, and text) are required for this node to function. The text input supports multiline text and dynamic prompts for flexible text processing. This node is marked as experimental in the source code.

Outputs

Output Name	Description	Data Type
`CONDITIONING`	Enhanced conditioning data with the added controlnet cross-attention parameters (`cross_attn_controlnet` and `pooled_output_controlnet`) derived from the CLIP text encoding	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): 95a798684ca8734bfff53c7b979b320f6834dc1a9553163d0e567243761000f1