Image
Creating one image from a prompt, with optional visual references.
Read as Markdown ↗Use it for: Creating one image from a prompt, with optional visual references.
What it can do: Generate from text alone, transform an existing image, use an uploaded/library image as a reference, or combine several references when the selected model supports them.
Inputs: Connected text plus the local prompt; compatible image references from Image, media assets, Character, Upscale, or Remove Background.
Output: One generated image. Wait for generation and image loading to finish before using it downstream.
Try it: Add Image, enter the courier prompt, select model/output settings, check the displayed Gems, and generate. Inspect the image before continuing.
Image input examples
| Task | How to set it up |
|---|---|
| Text to image | Write in Image directly, or connect Text to its prompt |
| Image to image | Connect a completed Image to another Image; describe the change |
| Asset reference | Use Image to image to choose an owned library image or upload |
| Multiple references | Connect two ready image sources; name their roles in the prompt |
For the multiple-reference example, use a character portrait and a harbor background: “Place the character from the first reference on the street in the second reference. Keep the yellow raincoat.” Check reference thumbnails and order before generating. Reference fidelity varies with the model and prompt.
Image output examples
The same image can supply another Image, Video, Text, Upscale, Remove Background, Script, Storyboard, or Motion Video. See Image outputs for the three most useful branches.
Remember: Reference counts and settings vary by model. Changing models can make existing inputs unsupported. Canvas image nodes create one image per run.
Watch the setup
These short demos show setup or existing results. Generation and analysis are started separately after you review the inputs and displayed price or allowance.