Nodes/Universal Input Hub by Steve Lasmin/Universal Input Hub by Steve Lasmin
ComfyUI Node

Universal Input Hub by Steve Lasmin

A ComfyUI node in utils with 15 inputs and 20 outputs.

By Eklipsis·Created 3 months ago·Updated 3 months ago· 0
Universal Input Hub by Steve Lasmin
  • model
  • clip
  • width
  • height
  • preview
  • seed
  • noise
  • steps
  • cfg
  • sampler_name
  • sampler
  • scheduler_name
  • sigmas
  • flux_sigmas
  • model
  • prompt
  • prompt_with_triggers
  • negative_prompt
  • trigger_words
  • positive
  • negative
  • help
max_size1024
aspect_ratiocustom
width_ratio2
height_ratio3
multiplier32
seed777
steps20
cfg4.0
samplereuler
schedulernormal
prompt
trigger_words
negative_prompt
Categoryutils

Inputs (15)

NameTypeDefaultDescription
max_sizeINT102432–4096Maximum dimension in pixels. The longest side of the image will match this value.
aspect_ratioCOMBOcustomSelect a preset aspect ratio, or choose 'custom' to define your own via width_ratio and height_ratio.
width_ratioINT21–999999Custom width component of the aspect ratio. Only used when aspect_ratio is set to 'custom'.
height_ratioINT31–999999Custom height component of the aspect ratio. Only used when aspect_ratio is set to 'custom'.
multiplierCOMBO32Round width and height to be divisible by this value. Use 64 for SD 1.5, 32 for SDXL, 16 for Flux.
seedINT7770–18446744073709550000Random seed for noise generation. Use 'control after generate' to randomize, increment, or keep fixed.
stepsINT201–10000Number of sampling steps. Higher values = more detail but slower generation.
cfgFLOAT4.00–100Classifier-Free Guidance scale. Higher = stronger prompt adherence. Flux works well at 1.0-4.0, SDXL at 7.0-8.0.
samplerCOMBOeulerSampling algorithm. 'euler' is fast and reliable. 'dpmpp_2m' is higher quality.
schedulerCOMBOnormalNoise schedule. 'normal' is standard. 'simple' works well with Flux. 'karras' for SDXL.
modelMODELThe diffusion model. Required for standard sigmas calculation and model passthrough.
clipCLIPThe CLIP text encoder. Required to convert text prompts into conditioning tensors.
promptSTRINGMain positive prompt describing what you want to generate.
trigger_wordsSTRINGTrigger words prepended to the prompt with '. ' separator. Leave empty to use prompt only.
negative_promptSTRINGNegative prompt describing what you want to exclude from the image.

Outputs (20)

NameTypeDescription
widthINT
heightINT
previewSTRING
seedINT
noiseNOISE
stepsINT
cfgFLOAT
sampler_nameSTRING
samplerSAMPLER
scheduler_nameSTRING
sigmasSIGMAS
flux_sigmasSIGMAS
modelMODEL
promptSTRING
prompt_with_triggersSTRING
negative_promptSTRING
trigger_wordsSTRING
positiveCONDITIONING
negativeCONDITIONING
helpSTRING