EasyKSampler (Inpainting)(samplerSimpleInpainting)

Description

The EasyKSampler (Inpainting) node is a purpose-built sampler from the ComfyUI-Easy-Use package for masked/region-based inpainting workflows. It accepts an EasyUse PIPE_LINE, optionally overrides the model and mask, and produces an updated pipeline, the resulting image, and the VAE used for the sampling process.

Inputs

pipe

The EasyUse PIPE_LINE containing the model, conditioning, latents, VAE, and other workflow state. The node reads the sampling inputs from this pipeline and returns an updated pipeline that can be chained into downstream EasyUse nodes.

grow_mask_by

An integer controlling how many pixels the inpaint mask is grown/expanded before sampling. Values can range from 0 to 64, with a default of 6. Growing the mask helps ensure the sampled region blends into the surrounding image context.

image_output

Controls what is done with the output image after sampling.

• Hide — Does not show or save the output image.
• Preview — Shows the image in the UI preview.
• Save — Saves the image with the save_prefix.
• Hide&Save — Saves the image without previewing it.
• Sender — Sends the image to a workflow link specified by link_id.
• Sender&Save — Sends the image to a workflow link and also saves it.

Default: Preview

link_id

An integer link ID used when image_output is Sender or Sender&Save. This determines where the generated image is sent. It has no effect for the other image output modes.

save_prefix

The filename prefix used when saving the output image. This is only relevant when image_output includes Save.

Default: ComfyUI

additional

The auxiliary inpainting strategy to use. This changes how the mask and conditioning are applied during sampling.

• None — Standard inpainting using the supplied mask.
• InpaintModelCond — Uses the model's native inpainting conditioning path.
• Differential Diffusion — Uses differential diffusion, treating mask values as per-pixel denoising strengths.
• Fooocus Inpaint — Applies the Fooocus inpainting strategy.
• Fooocus Inpaint + DD — Combines Fooocus inpainting with Differential Diffusion.
• Brushnet Random — Uses Brushnet with a randomly generated mask guidance.
• Brushnet Random + DD — Combines Brushnet Random with Differential Diffusion.
• Brushnet Segmentation — Uses Brushnet with segmentation-based guidance.
• Brushnet Segmentation + DD — Combines Brushnet Segmentation with Differential Diffusion.

Default: None

model

Optional model override. If supplied, this model is used for sampling instead of the model in pipe.

mask

Optional mask override. If supplied, this mask is used for the inpainting region instead of the mask in pipe.

Outputs

PIPE_LINE

The updated EasyUse pipeline, containing the sampling state after inpainting. This can be passed to another EasyUse node to continue the workflow.

IMAGE

The final output image after inpainting sampling.

VAE

The VAE associated with the inpainting process. This is useful for downstream encoding/decoding operations.

Usage Notes

This node is designed for inpainting workflows that already use an EasyUse pipeline. Typical usage connects a pipeline from an EasyUse loader/sampler chain, optionally provides a mask, and uses the returned pipeline for further processing.

When model or mask is not connected, the node falls back to the model or mask stored in the pipe. The grow_mask_by setting can help avoid hard mask edges by expanding the inpaint region before sampling.

For Brushnet and Differential Diffusion options, make sure the model and pipeline state are compatible with the selected additional strategy. The + DD variants combine the base strategy with Differential Diffusion-style denoising strength handling.