This workflow turns a text prompt into four high-quality, stylistically varied images using the Krea v2 image generation service. At its core is the Krea2ImageNode, which sends your prompt (and any optional parameters the node exposes) to the Krea API, returning a batch of images in one run. The SaveImage node then writes those images to your ComfyUI output directory with your chosen filename prefix.

Technically, ComfyUI orchestrates the request/response cycle while compute happens remotely on Krea. You supply the prompt in Krea2ImageNode and, if available in your installation, can set options like seed (for reproducibility), aspect/size, and the number of images. The workflow is preconfigured to produce four outputs for creative variety. The SaveImage node receives the image batch directly from Krea2ImageNode and saves each item as a separate file, which makes iterating and curating results straightforward.

Because generation runs in the cloud, this setup is fast to iterate on without a local GPU. It’s ideal for quickly exploring multiple visual directions from a single prompt, then narrowing in by locking a seed (if exposed) or refining the prompt. The minimal node graph—Krea2ImageNode plus SaveImage—keeps the process simple while still letting you manage output naming and batching inside ComfyUI.