Extract text from an image with OCR, then generate a new image from that text via an LLM-crafted prompt
Studio-style product image generated from the OCR'd content