We are Tuttebelle Fashion, a Dutch women’s fashion retailer. We are building an automated workflow with n8n and would like to use PhotoRoom for product recoloring.
Our use case is:
A1.jpg = our model wearing a blouse in one colour
B1.jpg, B2.jpg, B3.jpg, etc. = photographs of the same blouse in the actual other colours
We want to automatically generate:
A1 + B1 → same A1 photo, blouse recolored to B1 A1 + B2 → same A1 photo, blouse recolored to B2
etc.
It is important that face, pose, body, garment shape, texture, folds, lighting and background stay unchanged. Only the garment colour should change.
We have three questions:
Can the PhotoRoom API use a second image as the actual colour reference for Recolor?
Is the Recolor functionality available through the API so we can automate it with n8n?
Can PhotoRoom automatically detect and recolor only the garment being worn, without manually masking/selecting it for every image?
If useful, we can upload one model image plus several real colour-reference images to demonstrate the exact workflow.
Yes, you can automate this with Photoroom’s Image Editing API, but there’s an important limitation with the color reference.
The API currently can’t take B1.jpg as a second image and use its garment color directly as the Recolor reference. Recolor works from a specified target color rather than a second reference image. For your workflow, you’d need to extract the target color (for example, a HEX value) from B1/B2/B3 first, then use that color when processing A1.
You don’t necessarily need to create a manual mask for every image. You can instruct the edit to target the garment (for example, the blouse) while keeping the rest of the image unchanged. However, because this involves AI-based editing, I wouldn’t expect pixel-perfect preservation of the face, pose, garment texture/folds, lighting, etc. in every generation.
So the closest automated workflow today would be:
B1.jpg → extract target color → A1.jpg + target color → Photoroom Image Editing API → recolored A1
