Skip to main content

Managing Product Images

Product images are a critical input for AI enrichment in Merchkit. The Images page under Catalog in the sidebar is where you manage your product image library.

What the Image Library Contains

The Images page displays all images associated with your products. This is your centralized library for product photos, lifestyle shots, and any visual content that feeds into Merchkit’s enrichment engine.

How Images Relate to Products

Each product can have one or more images. When you import products, you can map a column in your CSV to the product’s image field. Images are indexed and organized by product SKU, making it easy to match images to the right products.

Using Images for AI Enrichment

Images are powerful inputs for AI-powered attribute generation. When you set up an attribute in Merchkit, you can tell the AI which images to analyze to generate or enrich that attribute.

Image Source in Attribute Configuration

For a prompt attribute that analyzes photos, choose images under Product Data, then select the relevant Image Source. This provides visual context for generating a value; it is separate from choosing existing image references from a parent’s gallery.

Select existing photos from a parent product

Suppose a chair has oak and walnut variants, while the main product holds all the photos. An AI-generated reference can select existing photos for each variant from that parent’s gallery. This selects links to images; it does not generate new photos or automatically inherit a gallery. Before starting:
  • Link each variant to its immediate parent product.
  • Make sure the parent’s image list already contains the candidate photos.
  • Have a product Reference or List Reference attribute targeting images, and permission to manage attributes (attributes.manage) and run generation (attributes.generate). This is different from the Image field type used to create new images.
  1. Open the destination reference attribute’s Configure Attribute panel.
  2. Enable Use AI and confirm that the reference targets images.
  3. Under Select From, choose the image list or lists that hold the parent’s candidate photos. This control appears for supported image and source references.
  4. After selecting a list, set Sourced from to Parent product. The default is This product.
  5. Write instructions describing which existing images to select, and include the variant information needed to distinguish them. For example: select photos matching the variant’s finish; do not select a walnut photo for an oak variant.
  6. Click Save changes, then run reference generation on a small selection of variants using your normal enrichment workflow.
  7. Review the selected images and any recorded rationale or confidence before exporting them.
The parent setting uses the immediate parent’s selected lists. It does not combine them with the child’s lists or search further ancestors. A missing parent or empty parent list supplies no candidates; adding the setting alone does not create or copy images. AI selection still needs review and can return no match. The same Select From and Sourced from controls apply to supported source references. If Sourced from is missing, check the reference’s target type, Use AI setting and whether you selected a candidate list first.

Save image editor columns

Open a product’s image list editor, use its column visibility control to show the fields your team needs, then click Save view. This account-wide setting requires views.manage. Image and source editors have separate defaults; see Saving editor columns.

Image Field Type & AI-Generated Images

Some attributes use the Image field type. These attributes can generate brand-new images using Merchkit’s image generation model (powered by Nano Banana). Use cases for AI-generated images:
  • Dimension line drawings — generated from product silhouettes to show scale and sizing
  • Lifestyle mockups — product images placed in real-world contexts
  • Transparent backgrounds — clean product shots with standardized backgrounds
  • Variations — color, angle, or zoom variations derived from source images
When you configure an Image field type attribute, you’ll specify which source images the AI should use as reference input. [SCREENSHOT: Image field type with AI generation options]

Uploading and Managing Images

Import Images with Products

The most common way to add images is through CSV import. When importing products:
  1. Include an image URL column in your CSV
  2. During the import mapping step, map this column to the Image field
  3. Merchkit will fetch and index these images with your products

Upload Images Directly

You can also upload images directly to the Images page:
  1. Click + Upload Images
  2. Select image files from your computer
  3. Map or confirm which product each image belongs to
  4. Click Save
[SCREENSHOT: Direct image upload interface]

Image Best Practices

  • Use high-quality images (at least 500x500 pixels recommended)
  • Ensure images are in common formats (JPG, PNG, WebP)
  • Use clear, well-lit product photos for best AI analysis
  • Include multiple angles per product if available — the AI will use these for more comprehensive enrichment

Next Steps

Now that your images are organized, you’re ready to learn how to handle import errors and validation issues. Head to Import Validation: Understanding Errors & Warnings.