This all-in-one workflow brings together the official FLUX 3 Klein workflows for text-to-image generation, single-image editing, and multi-image reference editing. It also includes the officially recommended prompt enhancement feature, which you can turn on or off with a simple switch.
How to use:
Import the JSON into ComfyUI and install the required models and missing nodes.
Select a generation mode. For editing, upload your reference images and enable the slots you need (up to 10 images).
Enter your prompt, choose whether to enable prompt enhancement, and set the output aspect ratio and size.
Run with the preset settings: 16 steps, CFG 1, Euler / Simple.
For editing, turn off the custom aspect ratio option to follow Image 1’s proportions, or turn it on to choose a different ratio. Output size is controlled by the megapixel setting.
Model downloads:
https://huggingface.co/black-forest-labs/flux-3-image-klein-comfy
Try the online workflow:
https://www.runninghub.ai/zh-cn/post/2110978558508560385/?inviteCode=rh-v1443