Nodes/Comfyui-Z-Image-Utilities/Z-Image Integrated KSampler
ComfyUI Node

Z-Image Integrated KSampler

Z-Image Integrated KSampler - All-in-one generation with LLM prompt enhancement, sampling, AuraFlow, CFGNorm, and auto-save.

By Koko-boya·Created 9 months ago·Updated 5 months ago· 127
Z-Image Integrated KSampler
  • model
  • clip
  • vae
  • config
  • options
  • image_1
  • image_2
  • image_3
  • image_4
  • latent
  • images
  • latent
  • enhanced_prompt
  • debug_log
positive_prompt
negative_prompt
generation_modetext_to_image
width1024
height1024
seed0
steps20
cfg7.0
sampler_nameeuler
schedulernormal
denoise1.00
prompt_templatechinese
enable_prompt_enhancetrue
batch_size1
auraflow_shift0.00
cfg_norm_strength0.00
enable_clean_gpufalse
enable_clean_ramfalse
auto_save_folder
output_prefixz_image
custom_system_prompt
instructionDescribe the key features of the input image (color, shape, size, texture, objects, background), then explain how the user's text instruction should alter or modify the image. Generate a new image that meets the user's requirements while maintaining consistency with the original input where appropriate.
CategoryZ-Image

Inputs (32)

NameTypeDefaultDescription
modelMODELDiffusion model for image generation
clipCLIPCLIP model for text encoding
vaeVAEVAE model for latent encode/decode
configZIMAGE_CONFIGLLM configuration for prompt enhancement
positive_promptSTRINGPositive prompt - will be enhanced by LLM
negative_promptSTRINGNegative prompt for unwanted elements
generation_modeCOMBOtext_to_imageT2I: Generate from text. I2I: Edit/transform input images
widthINT102464–8192Output image width
heightINT102464–8192Output image height
seedINT00–18446744073709550000Random seed
stepsINT201–1000Sampling steps
cfgFLOAT7.00–100CFG scale
sampler_nameCOMBOeulerSampler algorithm
schedulerCOMBOnormalScheduler type
denoiseFLOAT1.000–1Denoise strength (1.0 = full, <1.0 for I2I)
optionsoptZIMAGE_OPTIONSAdvanced LLM options
image_1optIMAGEFirst image for I2I/vision enhancement
image_2optIMAGESecond image (optional)
image_3optIMAGEThird image (optional)
image_4optIMAGEFourth image (optional)
latentoptLATENTOptional latent input (for custom latent)
prompt_templateoptCOMBOchineseLLM prompt template language
enable_prompt_enhanceoptBOOLEANtrueEnable LLM prompt enhancement (disable to use prompt as-is)
batch_sizeoptINT11–16Number of images to generate
auraflow_shiftoptFLOAT0.000–100AuraFlow shift (0 = disabled)
cfg_norm_strengthoptFLOAT0.000–10CFGNorm strength (0 = disabled)
enable_clean_gpuoptBOOLEANfalseClean GPU memory before/after sampling
enable_clean_ramoptBOOLEANfalseClean RAM after completion
auto_save_folderoptSTRINGAuto-save folder (empty = disabled)
output_prefixoptSTRINGz_imageFilename prefix for auto-save
custom_system_promptoptSTRING
instructionoptSTRINGDescribe the key features of the input image (color, shape, size, texture, objects, background), then explain how the user's text instruction should alter or modify the image. Generate a new image that meets the user's requirements while maintaining consistency with the original input where appropriate.System instruction for the Vision-Language Model in I2I mode

Outputs (4)

NameTypeDescription
imagesIMAGE
latentLATENT
enhanced_promptSTRING
debug_logSTRING