LoraLoader - ComfyUI Built-in Node Documentation

Complete documentation for the LoraLoader node in ComfyUI. Learn its inputs, outputs, parameters and usage.

This node automatically detects models located in the LoRA folder (including subfolders) with the corresponding model path being ComfyUI\models\loras. For more information, please refer to Installing LoRA Models

The LoRA Loader node is primarily used to load LoRA models. You can think of LoRA models as filters that can give your images specific styles, content, and details:

• Apply specific artistic styles (like ink painting)
• Add characteristics of certain characters (like game characters)
• Add specific details to the image All of these can be achieved through LoRA.

If you need to load multiple LoRA models, you can directly chain multiple nodes together, as shown below:

Inputs

Parameter	Data Type	Description
model	MODEL	Typically used to connect to the base model
clip	CLIP	Typically used to connect to the CLIP model
lora_name	COMBO[STRING]	Select the name of the LoRA model to use
strength_model	FLOAT	Value range from -100.0 to 100.0, typically used between 0~1 for daily image generation. Higher values result in more pronounced model adjustment effects
strength_clip	FLOAT	Value range from -100.0 to 100.0, typically used between 0~1 for daily image generation. Higher values result in more pronounced model adjustment effects

Outputs

Parameter	Data Type	Description
model	MODEL	The model with LoRA adjustments applied
clip	CLIP	The CLIP instance with LoRA adjustments applied