comfyui-mudknight-utils

Anime-focused prompt node. Splices category fields into a single positive prompt. Lines commented with # are stripped. Supports <lora:file:strength> syntax.

AnimaPromptNode

ApplyControlNetPipe

Base generation node that creates images from either empty latent(for t2i) or an image or latent (for i2i and inpainting).Resolution can be used with scale to generate images above 1MP.

BaseNode

Used internally in detailer nodes for cropping from bbox.

BBoxInsetAndCrop

CharacterPresetNode

CharacterReplacementNode

This is a convenience node when using an upscaling model to be able to scale the image to a factor based on the original image.

CombinedUpscaleNode

Add lora if substring is in positive_text.

ConditionalLoraFullPipe

Creates an empty latent using specific dimensions.

CustomResolutionPipe

Crops, upscales, samples, downscales, and uncrops the detected bbox

DetailerNode

DetailerPipeNode

FullPipeIn

FullPipeOut

FullPipePack

Show the difference between 2 images, with optional cropping around the detected changed area.

ImageDifference

InpaintNode

InpaintPipeNode

Load checkpoint, seed, and loras into full pipe

LoaderFullPipe

LoadImageToLatentPipe

Strips LoRA syntax from the input string, applies the LoRAs to model and clip, and outputs the cleaned string.

LoraExtractNode

Apply a LoRA to the model and CLIP in a full pipe.

LoraFullPipe

MaskDetailerNode

MaskDetailerPipeNode

MetaEditNode

ModelPresetNode

MultiStringConditioning

Detail faces and then eyes within the face. FastDetailer nodes always scale regions to 1MP so this does the nested detailing on the scaled image instead of doing it on the final composited image.

NestedDetailerNode

NestedDetailerPipeNode

Denoises an image using OpenCV's edge-preserving filter followed by a bilateral filter. Reduces noise while preserving edges and fine detail. GPU is strongly recommended for reasonable performance.

OpenCVDenoise

PipeMetaEditNode

Prompt node that conditions positive and negative text. Lines commented with # are stripped. Accepts LoRA syntax with <lora:file.safetensors:strength>. Model and CLIP strength can both be specified with another colon. Uses quality tags, embeddings, styles, characters, and tags from Preset Manager.

PromptConditioningNode

Save images from full pipe or optional image input.Optionally saves ComfyUI and A1111 metadata.ComfyUI samplers are mapped to their A1111 equivalents.

SaveFullPipe

SaveGif

Simple prompt node that conditions positive and negative text. Lines commented with # are stripped. Accepts LoRA syntax with <lora:file.safetensors:strength>. Model and CLIP strength can both be specified with another colon. Negative to NegPip uses an empty negative prompt and moves negative tags to the positive prompt with inverted weights.

SimplePromptNode

Load separate diffusion model, CLIP, and VAE, apply AuraFlow sampling, and create full pipe

SplitLoaderFullPipe

StylePresetNode

TagPresetNode

Do an i2i upscale on an image, optionally with an upscale model to improve sharpness.

ComfyUI Extension: comfyui-mudknight-utils

Custom Nodes (39)

Run ComfyUI workflows without the setup