AI Analyze Image action in workflows

Workflows can now analyze an image with an AI vision model. The AI Analyze Image action under AI Actions takes a public image URL and a prompt, returns the model's answer as text, and passes it to the actions that follow.

Choose GPT-5.6 Luna or GPT-5.6 Tera, enter the image URL (which can come from an earlier step or a custom value), write the analysis prompt, and set the detail level to Auto, High, or Low. High is more accurate for small text; Low is faster. Use it to read text from receipts, invoices, screenshots, and documents, or to get a written description of an image's contents, then save the result to a custom field, send it in an email or SMS, pass it to a webhook, or send it to another AI action.

This is a premium action billed per execution. The image URL must be public and point directly to the image file, not to a page that contains the image. Files in Media Storage are already public.

See Analyze an image with the AI Analyze Image action for full steps.