Why this matters
Image-to-image pipelines let you turn a single product photo into an entire visual asset bank — product photography, UGC-style images, CGI, and more — all while keeping typography, bottle form, and color palette consistent across every output.
What you'll build
A basic image-to-image pipeline in ImagineArt that takes one product image and generates multiple on-brand variations using a connected Copilot node and image node.
Step-by-step
Start with your reference setup. Import your product image into ImagineArt and connect it to the Copilot node. The Copilot node's system prompt should already define the typography, product form, and color palette so consistency carries through every generation.
Add an image node. Right-click on the workspace and select Image Node.
Choose your model. In the right-hand panel, pick from the available models based on your output needs — options include GPT Image 2, Nano Banana 2, and Nano Banana Pro. For this example, use Nano Banana 2.
Set your output properties.
Aspect ratio: 4:5
Resolution: 4K
Connect the nodes. Link the Copilot node to the image node, and attach the reference product image to the image node.
Write your prompt in the Copilot node. Example used in this lesson:
"Place my product on a white vanity and make sure that the composition is dynamic where the product is not necessarily centralized."
Select your generation count. Click and drag to select both the Copilot node and the image node. In the bottom-right corner, set the number of runs — three generations is a good starting point for comparing variations.
Check your credit cost. Hover over Run Selected to see how many credits the generation will consume.
Run the pipeline. Select Run. After generation completes, you'll have multiple image variations to choose from — in this case, three.
Takeaway
A connected Copilot + image node pipeline is the foundation for scalable, on-brand visual generation — one reference image and one prompt can produce multiple usable outputs in a single run.