A filled image workflow recipe recording the exact model files, workflow, prompt, seed and output for one result.
About
An image you liked is hard to repeat if the model version, companion files and seed were never written down. This workflow recipe keeps them together with the workflow, so the work is easier to repeat and share.
You record the model family and exact variant, model-version URL and digest, workflow source, ComfyUI version, generation weights, text encoder, VAE, any optional LoRA with its declared compatibility, model license sources, the positive prompt, seed, resolution and batch size, the prompt ID and the output. The notes tell you to start with Krea 2 Turbo or your exact FLUX version's matching template, export the working graph in API format, and record its actual prompt and seed bindings before changing them.
The small SD1.x graph is an optional structure exercise, not the main lesson's model. You need a text editor to fill it in. Provider accounts, model weights, GPU compute and hosting are not included.
- ComfyUI
Requirements
- Use with
- A text editor
- Operating systems
- windows, macos, linux
- Network