Skip to lesson content
Learn and help
Open editor

Edit images in the AI Image panel

Last updated: October 23, 2026 · 4 min read

Add a shadow, a new background, a light fix, a model fitting or any change you describe to the selected image.

The AI Image panel on the right rail never writes a result anywhere by itself. Results collect in the panel; adding one to the canvas, putting it in place of the selected image or continuing from it is your call. To generate a new image from scratch, use AI Image in the left panel.

Follow these steps

  1. 1

    Select the image and open the panel

    CanvasAI Image

    Select an image on the canvas and open the AI Image tab on the right rail. For an image on a product card, use the Open in AI Image link in Properties. The Working image area at the top shows which image you are working on; with nothing selected, use Pick from uploads to choose an image you uploaded.

  2. 2

    Choose a tool

    AI ImageTools

    Tools are grouped as Product photo, Fashion, Combine and mockup, and Retouch. Hover a tile to open an info card that shows what the tool does with a before and after example, along with its credit cost. Click a tile to open only that tool’s settings; use the Tools button to go back to the list.

  3. 3

    Complete the settings

    AI ImageToolsAI model

    For tools with a model choice, pick Nano Banana 2, GPT Image 2 or PhotoRoom from the AI model list; each option shows its credit cost. For tools that need an instruction, describe what should happen in one clear sentence. Dress a model needs a model photo and Combine needs at least one reference. If something is missing, the Generate button stays disabled and tells you what is missing.

    Only write the change you want. The model does not know about the previous run, so a sentence like “you forgot this, do it too” does not work; write the missing change directly.

  4. 4

    Generate

    AI ImageGenerate

    Press Generate. The result appears under Results. If the provider’s safety filter blocks a result, the panel shows a button to send the same request with another model.

  5. 5

    Use the result

    AI ImageResults

    Add to canvas adds the result as a new block in the image’s own proportions. Replace puts the result in place of the selected image on the canvas. Continue with this leaves the canvas untouched: it makes the result the working image and clears the instruction, so your next edit is made on it. Back to selection returns to the original image. Regenerate runs the same tool with the same settings again and uses credits again.

Tools

Each row is one tool in the panel. The Documentation link on a tool’s info card brings you straight to its row.

ToolWhere?What it does
Edit with AIAI ImageChanges the image according to your instruction: colors, background, items to add or remove. You choose the AI model.
ShadowAI ImageAdds a Soft, Hard or Floating shadow to the product. If the background has not been removed, it removes it first.
Background colorAI ImageRemoves the background and places the product on the solid color you pick.
Place in sceneAI ImageKeeps the product as it is and builds the scene you describe around it, for example a marble counter in morning light.
Light and colorAI ImageFixes exposure and color casts without touching the framing or the product. An optional note lets you steer the tone.
EnhanceAI ImageAutomatically sharpens and improves a product photo; the framing is kept.
Dress a modelAI ImageDresses the model photo you choose in the selected product. The model’s face, pose and background are kept.
Virtual modelAI ImageShows the garment on a model created by AI. An optional note lets you describe the rest of the outfit.
Ghost mannequinAI ImageShows the garment as if worn by an invisible mannequin, hollow and three-dimensional.
CombineAI ImageTakes what you want from up to three reference images and builds it on the working image. A reference is never pasted in as a separate card or a second picture.
MockupAI ImageShows the design on a t-shirt, cap, mug, tote bag, phone screen or laptop screen. If you pick your own blank product photo, the design is printed on it.
Magic eraserAI ImageErases the area you paint with the brush on the canvas. It only works on an image block on the canvas whose background has not been removed; see the Magic eraser lesson for the steps.
Remove textAI ImageRemoves text and watermarks from the image. Choose All, Watermark/added or Natural (signs).

Result

The result you like is on the canvas or in place of the selected image. The original image stays as it is unless you replace it.

Troubleshooting

The Generate button is disabled
The note above the button tells you what is missing: an instruction, a model photo or a reference. If you do not have enough credits, the run does not start and the panel tells you.
The result was blocked by a safety filter
Providers sometimes reject harmless images too. Use the button in the warning to send the same request with another model.

Related lessons