ComfyUI Workflow for Image Generation

• Needed Custom Node: RvTools v2

• This workflow is build upon FLUXinator

  • it contains Face Swap and Restore

  • more Detailers

  • more Filters

  • and outpaint wich seems to create better results than pulid especially in the variation of the face positions.

Outpaint

to use this feature you need the Load Image Group, differential diffusion (model patcher), flux guidance is set to 50, the portrait lora, a flux fill checkpoint, and you have to change the switch in the sampler group to yes for outpaint. i would recommend to use the face detailer to fix bad skin.

• i've found the workflow from Sebastian Kamph and implemented it in this one (basicly added 3 nodes to make it work in my workflow):

• https://www.patreon.com/posts/ace-character-1-121116973

Face Swap

if you have ReActor already installed there are no updates anymore but a new version 0.6 that contains an nsfw filter. if you want to use it you have to install the new version. both are able to run side by side as far as i can tell.

with the latest comfyui version (pytorch 2.7) the masking helper is not working, bypass it in the fast bypasser if you have issues