Skip to content

Generate or edit images with AI

Generate new images or modify existing ones using AI image models with the KNIME AI Extension.

When this is useful

Use image generation or editing when you want to:

  • Produce visuals automatically from structured or unstructured data
  • Create illustrations or mockups as part of an automated workflow
  • Include generated images in Data Apps

If you only need text-based output, see Prompt a model instead.

Generate or edit images in KNIME

Prerequisites
  • A provider and credentials configured
    (see Providers: Reference table)
  • An image-capable model available via an LLM Selector
  • KNIME AI Extensions installed

Create the image prompt

Image prompts are plain text instructions that describe what the model should generate or modify.

You can create prompts dynamically using nodes such as:

Example prompt:

txt
Create a realistic illustration of a modern coffee shop interior.

Each input row represents a separate image generation request.

(Optional) Provide an input image

To edit an existing image, pass it as input to the image generation node.

  1. Click the plus (+) icon on the node to add an additional table input port.
  2. Connect the table containing the image you want to modify to the newly added table input port of the image generation node.

Generate or edit the image

Use the appropriate Image Generator node:

Each input row produces one or more images as output.

Store or display the result

The generated or edited images are returned as image outputs and can be:

  • written to disk
  • displayed in views
  • passed to downstream workflows
  • embedded in reports or applications

Result

After execution, the workflow produces one or more images generated or edited according to the provided prompt and inputs.

To see a complete example of image generation in action, check out this workflow on the KNIME Hub:

➡️ Generate product image variants with AI

Next steps

Combine image generation with text prompting for richer workflows