Image To Image

Overview

The Image To Image node transforms an input image by applying a style transfer, inpainting, or prompt-guided diffusion pass, producing a modified image as output. Use it to restyle contender outputs, fill masked regions, or apply visual transformations within a battle workflow. The node requires a valid image input (URL or base64) and a text prompt describing the desired transformation; an API credential for the underlying image model must be configured. On failure, execution routes to the error output port with a structured error payload.

Configuration

Field	Type	Required	Description
model	enum	Yes	The image generation model to use for the transformation (e.g. stable-diffusion-xl, dall-e-3, flux-dev).
prompt	string	Yes	Text prompt describing the desired output image or transformation style.
negative_prompt	string	No	Text describing features to suppress or avoid in the output image.
strength	number	No	How strongly the prompt overrides the input image. Range 0.0–1.0; lower values preserve more of the original (default: 0.75).
mask_image_key	string	No	Workflow data key for a binary mask image used in inpainting mode. When provided, only the masked region is transformed.
steps	number	No	Number of diffusion steps. Higher values improve quality at the cost of latency (default: 30).
credential_id	string	Yes	ID of the stored API credential for the selected image model provider.

Inputs

Port	Type	Description
image	string	Source image to transform, as a URL or base64-encoded data URI.
mask	string	Optional binary mask image (URL or base64) defining the inpainting region. Overrides mask_image_key in config when present.
prompt_override	string	Optional runtime prompt that overrides the static prompt set in config.

Outputs

Port	Type	Description
output	string	The transformed image as a base64-encoded data URI or URL, depending on provider response format.
metadata	object	Provider response metadata including model used, seed, steps, and finish reason.
error	object	Structured error payload emitted when the transformation fails (e.g. invalid image, credential error, timeout). Contains code and message fields.

Example

{
  "nodeType": "image_to_image",
  "config": {
    "model": "stable-diffusion-xl",
    "prompt": "oil painting style, warm golden tones, impressionist brushwork",
    "strength": 0.6,
    "steps": 30,
    "credential_id": "cred_sdxl_prod_01"
  }
}