Image Generation node

The Image Generation node creates images from a text prompt using a dedicated image model.

Handles

Handle	Purpose
prompt	The text prompt describing the image
images	Reference images for image-to-image / conditioning

The node stores no prompt internally — text and references arrive through the handles.

Configuration

• Model — the image model (e.g. openai/gpt-image-2.5-flare).
• Size — output dimensions.
• Aspect ratio — e.g. 16:9, 1:1.
• Number — 1–4 images.
• Seed — deterministic output.
• Quality — low / medium / high / auto. Defaults to medium for gpt-image models (significantly faster than the provider default).

Example — image-to-image

Wire an Upload File (image) into images, and a Prompt into prompt:

prompt: "Closeup of the subject's face, dramatic lighting"
images: <uploaded photo>

The generated image ({ dataUrl, mime }, or an array when n > 1) becomes the payload for downstream nodes.

Chaining

An Image Generation node’s output can feed another Image Generation node’s images handle — “generate a scene, then a closeup of it.”

Notes

• The card shows a full-height preview; click to enlarge in a lightbox.
• Every generation is saved to the node’s artifact history (downloadable).

Related

• Prompt — the text prompt.
• Video Generation — turn the image into a clip.