Simple ComfyUI SDXL / Pony / Illustrious / Flux workflow

ComfyUI & Workflows saved in each image

Each image created by ComfyUI includes the workflow inside the generated image (unless stripped). This means you can save any image generated in ComfyUI (such as the image samples from my models), and drag/drop them into ComfyUI for their complete workflow.

Complex Workflows

Since I use some niche nodes and setups, I do not recommend these workflows as a starting point, therefore I created these simple workflows you can find below.

ComfyUI Built-in Workflows

ComfyUI has built-in workflows under the Workflow > Browse Templates menu.

Simple Workflows

I created 2 simple workflows that work well with my models. They will of course also work with other models. You can find them in the Attachments in the RIGHT SIDE of the article, in image and json form.

SDXL / Pony / Illustrious Workflow

Download in the Attachments to the right, the Simple SDXL-PDXL-ILXL - LCM - Turbo.json file.

This uses the DMD2 LoRA to speed up the generation time. You can remove this LoRA, but you also need to edit the scheduler and sampling settings when you do. Carefully read the recommended settings on the model you are using for more instructions on this.

Is it still blurry/bad quality? Did you really read the model instructions carefully? Look at the other images.

Still bad? Really, read the instructions!

Flux Workflow

Download in the Attachments to the right, the Simple Flux Workflow.json file.

There are many different ways flux models can be loaded and used. This is the one I use. Any way should be fine really.

If you are unable to find the "Image Resize By Longer Side" node, you can find it in this node pack: https://github.com/jamesWalker55/comfyui-various

Alternatively, use a 2x upscaler instead, or resize the image without an upscaler.

Not Working?

If these workflows are too complex or not working for you, please start with the ComfyUI templates and explore from there. Try to add one piece at a time to make it similar to the workflows above.

If you need the models used by the workflows, follow the official guides at:

https://comfyanonymous.github.io/ComfyUI_examples/

clip_i + t5xxl_fp16 -- https://github.com/Comfy-Org/ComfyUI/discussions/4222

Also found clip_g -- https://huggingface.co/Comfy-Org/stable-diffusion-3.5-fp8/tree/main/text_encoders