@radnolan: One of the things that take you to power user in @Figma Weave is Text Iterators. Iterating with LLMs is all about running multiple prompts with the same references, then expanding & riffing from there. So connect an image with text on it to an image model. The prompt will guide it to change nothing but the poses. And the poses will be added through a variable. That variable is connected to a text iterator which splits a list of poses separated by an asterisk. Now, when you run the image model, it loops for each pose output in the list. But it’s so much more powerful than just iterating poses. See, you could take an image with text on it and get it translated and localized for any market with just a few more prompts. To do that, we send the original image to an Any LLM node which we prompt to transcribe the image, then run it through an iterator that translates each line in the ad. Now you COULD run that through an image model, but you can add one more prompt that explains which words should not be translated (like a product name or company name). Now, you can come back to this workflow, add new images, poses, and markets. Then you’ll get localized campaigns as scale. Save this for your next campaign! #figmaweaveambassador

Nolan

Open In TikTok:

Region: GB

Monday 31 August 2026 14:32:13 GMT