Qwen FaceSwap WF optimized for 16GB VRAM ~30-60 sec generations (NSFW Face Obstruction)

Based on NRDXs BFS Lora and Workflow

This workflow was created to supply yet another tool in your ComfyUI toolbox for making AI characters / influencers.

You generated the perfect fictional AI character, but her face is fucked up - everything else is perfect. You now have the opportunity to swap her with another gen where the face was perfect but everything else sucked. Saves time.

It should go without saying - but you guys are monkeys:

DO NOT UNDER ANY CIRCUMSTANCES USE THIS ON REAL PEOPLE!!!

I can and will not take responsibility for any of your mischievous endeavors.

---

Update 07/2026

You should use Flux 2 Klein and BFS Lora - much better, no plastic skin. There exist checkpoints that have the lora baked in as well, check your options and see what works best for your images and purposes.

---

QWEN Edit 2511 + BFS Lora optimized for 16GB VRAM ~30-60 sec generations.

• No more / less weird twirly zig zag pattern hair (depending on inputs)

• Good facial obstruction, works for BJ, Facial, Masks, Glasses, Blindfolds etc.

To achieve good face swaps with Facial Obstruction like objects, apparrel, fluids, etc. your reference images should have highest possible resolution and similar aspect.

Optional Upscale (CTRL+B bypass if you don't want it)

Steps: between 4-12, works good - I use 6

Just load workflow and install missing stuff.

https://civitai.com/models/2027766?modelVersionId=2556739

This should be good out of the box - everything else is ComfyUI basics which you will have to learn. No support given here. Figure it out. Not gatekeeping but it is imperative you figure out how this stuff works - it is very rewarding and opens doors that can otherwise not be opened for you - no one online can or will tell you what or where some doors are.