Skip to Content
Raydo

Image workflow nodes

Choose an image generation mode and map prompts, reference images, and outputs.

Upload image (image_input) selects a local picture and exposes imagePath, referenceImagePaths and references. Use Text to Image (text_to_image) for a prompt-led request, or Edit image (image_to_image) when the picture needs to be edited. Both generation modes execute image.generate.

  1. Select an available model for the chosen mode. Check the model’s supported image sizes and aspect ratios in the editor.
  2. Map the text into prompt. For an edit, map the upstream imagePath into referenceImagePath, or provide a compatible references array.
  3. Check the generated image in the run. Look for artifactId, artifactPath, or preview in the current output. Older definitions can expose fields such as imagePath or imageArtifact on a generation step, so confirm what the saved workflow returns before reusing its mapping.

If a reference is ignored, compare the resolved input with the selected mode and check that the source file is readable on the execution device. An explicit `references` array takes precedence over `referenceImagePath`, even when that array is empty. For project delivery, choose the reviewed image’s artifact ID.

On this page

Image workflow nodes