Carousel slide images from a main image

A webhook receives the ids of a carousel record in Airtable and sets its AI status to AI Writing.... The workflow downloads the carousel's main image and sends it to the OpenAI image edit endpoint five times, once with each slide prompt (slides 2 to 6) and the brand style saved on the record, so every slide keeps the main image's theme with no text.

Each new image is uploaded to Cloudinary, the five links are merged, and the record's carousel images field is filled with them and set to AI Complete. A few leftover nodes, an image analysis step and a split and loop, sit in the file but are not connected.