All products

Tag product photos in ImageKit with AI

Tag catalogue photos from an approved vocabulary, fill existing metadata fields and give uncertain quality checks their own outcome in the media library.

To tag product photos in ImageKit with AI, decide which labels your catalogue actually needs. A furniture shop might allow oak, walnut and painted finishes, alongside a separate flag for a photograph that cuts off the chair legs.

The ai-tasks skill helps the agent define those choices and what should happen after each result. It can write an existing custom field, apply several tags or route an uncertain quality check to manual review.

What it does

  • Tag clothing photographs with the product categories that appear in the shop’s filters.
  • Populate an existing field with a dominant colour drawn from approved values.
  • Check whether a product is fully visible and flag unclear results for a person.
  • Limit how many furniture categories can be added to a single room scene.
  • Save the task for a batch, an upload workflow or assets matching a chosen path.

Install and use

  1. Install the skill and connect an ImageKit account with DAM and AI Tasks access. Create any intended custom metadata field first.
  2. Ask, “Tag each bag by material from this list, and mark images with a partly hidden product for review.”
  3. Check a small set of results before applying the task across the collection. Adjust overlapping labels and the unknown branch where needed.

FAQ

Source

Made by ImageKit. Original skill: ai-tasks, published on GitHub under the MIT licence.

View on GitHub