Menu

AI Image

Generate an image from a prompt and save it to your Media library

The AI Image step sends a prompt to an AI image model and saves the picture it creates in your Media library. A later step can use it, for example a Posting step that publishes it. You use your own account with an AI provider, so you pay the provider directly and choose any image model it offers.

Connect an AI provider

AI Image uses the same OpenAI-compatible connection as AI Text (OpenAI, OpenRouter, or another server that speaks the OpenAI API). When you add the connection, you can fill the optional Default image model, for example gpt-image-1 on OpenAI or an image model ID from OpenRouter's model list. AI Image steps use it unless they pick another model. Leave it empty if you only use AI Text, or pick a model in each AI Image step.

Not every provider offers image generation. If your server has no image models, the step fails with the provider's message.

Add the AI Image step

  1. In the workflow editor, press +.
  2. Pick AI Image (under AI). It starts with one of your AI connections selected; the step's form lets you choose another.
  3. Fill in the fields below.

Every field accepts {{ <stepId>.<path> }} expressions, like other steps.

Field Notes
Connection required; the provider this step calls
Prompt required; describe the image, usually with data from earlier steps

The AI Image step's settings

Under Advanced (collapsed), all optional:

Field Notes
Model empty uses the connection's default image model
Size for example 1024x1024. Empty uses the provider's default. Allowed values depend on the model
Quality for example low, medium or high. Empty uses the provider's default. Allowed values depend on the model

Each run creates one image.

Output and failures

On success the image is a new file in your Media library and the step's output has:

  • fileId: the file's ID, for example {{ step1.fileId }}. Use it in a Posting step under Images from an expression
  • mediaType: for example image/png
  • name: the file name Alur gave it, like ai-image-1a2b3c4d.png
  • size: the file size in bytes
  • model: the model that was requested

A request that is rejected (a bad key, an unknown model, a prompt the provider refuses, no credit left), a reply with no image, or an image over the Media limit of 25 MB fails the step with a clear message, and the run shows it. Alur does not retry these, because each attempt would be billed. A request that takes longer than two minutes also fails without a retry: the provider may still finish and bill the image. Only a rate limit (HTTP 429) and network errors are retried automatically.

Alur does not guarantee that a call is made exactly once if something fails mid-way, and every image is billed by your provider, usually at a higher price than text. In a For each loop each item generates its own image, so a long list costs one image per item; the step's form shows a warning when it sits inside a For each.

Delete generated images you no longer need from Media.