Skip to content

creatify-batch: batch image generation

Ad creative (utility) Desktop

creatify-batch generates images in bulk via the Creatify Asset Generator (Nano Banana Pro). You give it a list of prompts, it sends them to Creatify, waits for the renders to finish, and downloads the results to your project folder.

Most of the time you won’t call it directly. Skills like meta-image-ad-generator and living-elements call it behind the scenes as part of their own workflow. It is the image-rendering engine that powers those skills.

  • You have a set of ad prompts written and just want to render them, without running the full ad-generator workflow.
  • You want to re-render a batch with tweaked prompts after a previous run.
  • You’re experimenting with prompt variations and want to see the results side by side.
  • A downstream skill ran into an error mid-batch and you want to re-run only the failed prompts.

For most use cases, you won’t reach for this directly. The ad skills handle it for you.

Desktop only

creatify-batch needs the Claude Code desktop app and your Creatify API keys set in the environment. The Cowork sandbox can’t reach outside image services, so this tool won’t run there. See API key setup to get your keys in place.

If you want to call it directly, just describe what you need:

“Generate these five ad images using creatify-batch: [paste your prompts]”

Or reference a prompt file you’ve already written:

“Run creatify-batch on my prompts file for AcmeCo and save the images to the project folder.”

If your prompts include product shots, tell it where to find the reference image:

“Run creatify-batch for AcmeCo, using the product photo at [CDN URL] as the reference image for every prompt.”

  • Creatify API keys set in your environment (CREATIFY_API_ID and CREATIFY_API_KEY). The script exits loudly with setup instructions if they’re missing.
  • A set of image prompts. You can describe them in plain language or point to a prompts file.
  • A product reference URL for any prompt that shows the brand’s product. Creatify needs a publicly accessible image URL to ground the render to the real product. Without one, it invents a generic lookalike that won’t match your brand.

Rendered image files downloaded to your project folder. You’ll see one file per prompt, ready to review and use in your ads or pages.

You: “Run creatify-batch for ExampleBrand. Here are four ad prompts. Use the product photo at https://cdn.examplebrand.com/product.png as the reference image.”

creatify-batch: (sends prompts to Creatify, polls until renders are complete, downloads results)

“Done. Four images saved to your project folder: ad-1.jpg, ad-2.jpg, ad-3.jpg, ad-4.jpg. Check them against the prompts before using.”

The renders typically take a minute or two per image. The skill waits automatically and notifies you when everything is ready.

  • You usually won’t need to call this directly. If you’re using meta-image-ad-generator or a similar skill, it’s already running creatify-batch for you as part of the workflow.
  • API keys are required, and the script is not quiet about it. If keys are missing, you’ll get a clear error message with setup instructions. See API key setup.
  • Product images need a real reference URL. If you skip the reference for a product shot, Creatify generates a plausible-looking but wrong product. Pass a public CDN link from the brand’s own site for accurate results.
  • Text in prompts often needs 2-3 renders to land correctly. AI image models can mangle exact wording. Budget for a re-run if your prompts include specific copy.

creatify-batch is called automatically by meta-image-ad-generator and living-elements. It is a sibling of image-sourcer, which finds and downloads existing real photos rather than generating new ones.