Skip to main content
Conditioning nodes convert text prompts into embeddings and manipulate how they guide the image generation process.

​
Text Encoding

​
CLIPTextEncode

Encodes a text prompt using a CLIP model into an embedding that guides the diffusion model. Category: conditioning
​
text
string
required
The text prompt to encode. Supports multiline and dynamic prompts.
​
clip
CLIP
required
The CLIP model used for encoding the text
Returns:
  • CONDITIONING: The encoded text embedding used to guide the diffusion model
Description: Encodes a text prompt using a CLIP model into an embedding that can be used to guide the diffusion model towards generating specific images. This is the primary way to create positive and negative prompts.

​
CLIPSetLastLayer

Controls which CLIP layer to use for text encoding. Category: conditioning
​
clip
CLIP
required
The CLIP model to modify
​
stop_at_clip_layer
int
default:"-1"
Which CLIP layer to stop at. Range: -24 to -1. -1 uses the final layer, -2 uses the second-to-last layer, etc.
Returns:
  • CLIP: The modified CLIP model
Description: Different CLIP layers capture different levels of detail. Earlier layers (like -2) sometimes produce better results for certain styles.

​
Conditioning Combination

​
ConditioningCombine

Combines two conditioning inputs by concatenating them. Category: conditioning
​
conditioning_1
CONDITIONING
required
First conditioning to combine
​
conditioning_2
CONDITIONING
required
Second conditioning to combine
Returns:
  • CONDITIONING: Combined conditioning
Description: Merges two conditionings together. Useful for applying multiple prompts or effects to a single generation.

​
ConditioningAverage

Blends two conditionings with weighted interpolation. Category: conditioning
​
conditioning_to
CONDITIONING
required
The target conditioning
​
conditioning_from
CONDITIONING
required
The source conditioning to blend in
​
conditioning_to_strength
float
default:"1.0"
Weight of the target conditioning. Range: 0.0 to 1.0
Returns:
  • CONDITIONING: Blended conditioning
Description: Creates a weighted blend between two conditionings. A strength of 1.0 uses only conditioning_to, 0.0 uses only conditioning_from, and 0.5 creates an equal blend.

​
ConditioningConcat

Concatenates conditioning tensors along the token dimension. Category: conditioning
​
conditioning_to
CONDITIONING
required
The target conditioning
​
conditioning_from
CONDITIONING
required
The conditioning to concatenate
Returns:
  • CONDITIONING: Concatenated conditioning

​
Spatial Conditioning

​
ConditioningSetArea

Applies conditioning to a specific rectangular area of the image. Category: conditioning
​
conditioning
CONDITIONING
required
The conditioning to apply to the area
​
width
int
default:"64"
Width of the area in pixels. Range: 64 to 16384, step 8
​
height
int
default:"64"
Height of the area in pixels. Range: 64 to 16384, step 8
​
x
int
default:"0"
X coordinate of the area. Range: 0 to 16384, step 8
​
y
int
default:"0"
Y coordinate of the area. Range: 0 to 16384, step 8
​
strength
float
default:"1.0"
Strength of the conditioning in this area. Range: 0.0 to 10.0
Returns:
  • CONDITIONING: Area-modified conditioning
Description: Applies a prompt to a specific region of the image. Useful for composing scenes with different prompts in different areas.

​
ConditioningSetAreaPercentage

Applies conditioning to an area specified by percentage of image dimensions. Category: conditioning
​
conditioning
CONDITIONING
required
The conditioning to apply
​
width
float
default:"1.0"
Width as percentage. Range: 0.0 to 1.0
​
height
float
default:"1.0"
Height as percentage. Range: 0.0 to 1.0
​
x
float
default:"0.0"
X position as percentage. Range: 0.0 to 1.0
​
y
float
default:"0.0"
Y position as percentage. Range: 0.0 to 1.0
​
strength
float
default:"1.0"
Conditioning strength. Range: 0.0 to 10.0
Returns:
  • CONDITIONING: Area-modified conditioning

​
ConditioningSetAreaStrength

Sets the strength of an existing area conditioning. Category: conditioning
​
conditioning
CONDITIONING
required
The conditioning to modify
​
strength
float
default:"1.0"
New strength value. Range: 0.0 to 10.0
Returns:
  • CONDITIONING: Strength-modified conditioning

​
ConditioningSetMask

Applies conditioning using a mask to define the region. Category: conditioning
​
conditioning
CONDITIONING
required
The conditioning to apply
​
mask
MASK
required
The mask defining where to apply the conditioning
​
strength
float
default:"1.0"
Conditioning strength. Range: 0.0 to 10.0
​
set_cond_area
string
default:"default"
How to handle the conditioning area. Options: default, mask bounds
Returns:
  • CONDITIONING: Mask-modified conditioning

​
Temporal Conditioning

​
ConditioningSetTimestepRange

Applies conditioning only during specific timesteps of the sampling process. Category: advanced/conditioning
​
conditioning
CONDITIONING
required
The conditioning to modify
​
start
float
default:"0.0"
Start of the timestep range. Range: 0.0 to 1.0 (0.0 = beginning of sampling)
​
end
float
default:"1.0"
End of the timestep range. Range: 0.0 to 1.0 (1.0 = end of sampling)
Returns:
  • CONDITIONING: Timestep-limited conditioning
Description: Useful for prompt scheduling, where different prompts apply at different stages of the generation process.

​
Advanced Conditioning

​
ConditioningZeroOut

Zeros out the conditioning, creating a null/empty conditioning. Category: advanced/conditioning
​
conditioning
CONDITIONING
required
The conditioning to zero out
Returns:
  • CONDITIONING: Zeroed conditioning
Description: Creates an empty conditioning. Useful for unconditional generation or as a baseline for certain advanced techniques.

​
ControlNet Conditioning

​
ControlNetApplyAdvanced

Applies ControlNet guidance to positive and negative conditioning. Category: conditioning/controlnet
​
positive
CONDITIONING
required
Positive conditioning to apply ControlNet to
​
negative
CONDITIONING
required
Negative conditioning to apply ControlNet to
​
control_net
CONTROL_NET
required
The ControlNet model to use
​
image
IMAGE
required
The control image (e.g., edge map, depth map, pose)
​
strength
float
default:"1.0"
ControlNet influence strength. Range: 0.0 to 10.0
​
start_percent
float
default:"0.0"
When to start applying ControlNet. Range: 0.0 to 1.0
​
end_percent
float
default:"1.0"
When to stop applying ControlNet. Range: 0.0 to 1.0
​
vae
VAE
Optional VAE for processing the control image
Returns:
  • positive: ControlNet-modified positive conditioning
  • negative: ControlNet-modified negative conditioning

​
Vision Conditioning

​
CLIPVisionEncode

Encodes an image using CLIP Vision for image-based conditioning. Category: conditioning
​
clip_vision
CLIP_VISION
required
The CLIP Vision model to use
​
image
IMAGE
required
The image to encode
​
crop
string
default:"center"
How to crop the image. Options: center, none
Returns:
  • CLIP_VISION_OUTPUT: Encoded image features

​
unCLIPConditioning

Applies unCLIP image conditioning to text conditioning. Category: conditioning
​
conditioning
CONDITIONING
required
The text conditioning to enhance
​
clip_vision_output
CLIP_VISION_OUTPUT
required
The encoded image features
​
strength
float
default:"1.0"
Image conditioning strength. Range: -10.0 to 10.0
​
noise_augmentation
float
default:"0.0"
Noise added to image embedding. Range: 0.0 to 1.0
Returns:
  • CONDITIONING: Image-conditioned text conditioning

​
StyleModelApply

Applies style from a reference image using a style model. Category: conditioning/style_model
​
conditioning
CONDITIONING
required
The conditioning to apply style to
​
style_model
STYLE_MODEL
required
The style model to use
​
clip_vision_output
CLIP_VISION_OUTPUT
required
Encoded features from the style reference image
​
strength
float
default:"1.0"
Style strength. Range: 0.0 to 10.0
​
strength_type
string
required
How to apply strength. Options: multiply, attn_bias
Returns:
  • CONDITIONING: Style-modified conditioning

​
Inpainting Conditioning

​
InpaintModelConditioning

Prepares conditioning for inpainting models. Category: conditioning/inpaint
​
positive
CONDITIONING
required
Positive conditioning
​
negative
CONDITIONING
required
Negative conditioning
​
vae
VAE
required
VAE for encoding the image
​
pixels
IMAGE
required
The image to inpaint
​
mask
MASK
required
The inpainting mask
​
noise_mask
bool
default:"true"
Add a noise mask to limit sampling to the masked area
Returns:
  • positive: Inpaint-prepared positive conditioning
  • negative: Inpaint-prepared negative conditioning
  • latent: Prepared latent with mask

Build docs developers (and LLMs) love

Get started for freeTalk to us