Producing an ad image

Format

You choose the size the output is produced at: square, portrait, story, landscape or small. Pick it for where the ad will run — square or portrait for feeds, story for stories.

Producing from a template

With a template the scene arrives ready and only the text slots are asked of you. Leave a slot empty and the AI writes it from your product. The template’s typeface, palette and output size are listed on its card; the palette can be changed from the workbench.

The “extra note” field does not change the scene — it informs the analysis: campaign context, tone, things to avoid. “Extra prompt” is detail you write on top of the scene the template builds.

Variants

You can produce up to four variants in one run. Each variant spends its own credits; if you are testing A against B, starting with two is usually enough.

Models

When you want the product shown on or in someone’s hands, you pick a model: gender, age, body type and scene (studio white, studio neutral, street, indoor loft, outdoors). Models are not a separate tab but a shared axis of the prompt and product modes — the same cast opens whichever mode you are in.

If a produced image contains a spelling error, replacing the problematic word with a synonym usually works faster than rerunning the same prompt — especially with long words.