AI Image Generation action in workflows
Workflows can now generate an image from a prompt. The AI Image Generation action under AI Actions picks a model, creates the image, stores it at a public URL, and returns that URL as text for the actions that follow.
Choose from four models ā GPT Image 2, Gemini 3 Pro Image (Nano Banana Pro), Gemini 3.1 Flash Image (Nano Banana 2), or Gemini 2.5 Flash Image (Nano Banana) ā set the quality to Auto, High, Medium, or Low, and write the prompt yourself or insert a custom value so it comes from a contact field, a webhook payload, or an earlier action. Enhance Prompt expands a short prompt, and templates let you start from Product Mockups, Social Post, Promo Flyer, or Lifestyle. Additional Settings control size, background, output format, Design Kit, and Brand Voice. Each image goes to Media Storage and returns an Image URL variable.
This is a premium action billed per execution. Your cost depends on the model, quality, size, and any additional context from a Design Kit or Brand Voice. The generated image is stored at a public URL, so anyone with the link can open it without logging in.
See Generate an image with the AI Image Generation action for full steps.