AI Cutout and White Background Starter
Prepare a single product reference, remove its background, and inspect the delivered image.
The ai_cutout_white_background template prepares a brief, calls image.background_remove, and passes its image, artifact path, and preview URI to the delivery node. Start with a single product photo and a dedicated output directory.
| Input | How the draft uses it |
|---|---|
sourceImageRefs, prompt | Required text. The draft prepares the source reference and a final prompt for the cutout node. |
backgroundMode, outputDirectory | Required text. Use white for a white background. Confirm the selected output directory is writable. |
namingPattern, qualityChecklist | Optional naming preference and review criteria. Check the actual output name for the selected adapter. |
The draft maps sourceImageRefs to imageRefs. Verify the selected node's inputs in the editor. If it requests a recognized reference field, map the single image to referenceImagePath or to an explicit references array for image editing. For a local adapter that expects imagePath, map the file path there.
{
"sourceImageRefs": "/path/to/test-product.png",
"prompt": "Preserve the product's small details and clean the edges",
"backgroundMode": "white",
"outputDirectory": "/path/to/raydo-cutout-test",
"namingPattern": "product-cutout-test",
"qualityChecklist": "No missing details, no halo, clean white background"
}Run with a model or adapter ready for image editing or cutout. In the result, check the subject edges, the background, the artifact path, and the preview. If recording project delivery, use the supported Delivery node settings and map the media artifact ID from the successful cutout step. For transparent output, inspect the exported file for an alpha channel, since provider support can vary. Use Product Image Batch as a reference for three separate image branches, and replace its generation actions if you need batch cutout.
See Image nodes and the Images workspace.