Combining Ideogram-4 with QwenVL in ComfyUI workflow

View on Reddit

A Reddit user shares a workflow that uses QwenVL to generate prompts from images, feeding them into ComfyUI's Ideogram-4 node without the template prompt enhancer. The approach enables image-to-image generation via an intermediate prompt step.