Mindket

BFS Best Face Swap

Newversion 1 · updated

Overview

LoRA add-ons that swap a face, a head or a whole person in an image, built for Qwen Image Edit, Flux 2 Klein and Krea 2. For image editors who have the right to use the faces involved.

BFS Best Face Swap is a kit in the Productivity category, listed by Mindket. It is free. This is version 1, last updated on . Download it on this page, a zip of its files. Nothing in it runs by itself.

At a glance

You need
  • Python on your computer
  • The huggingface_hub package
  • The base model that matches the LoRA you choose
  • A workflow for that base model
Runs in
  • Qwen Image Edit 2509 and 2511 workflows
  • Flux 2 Klein 4b and 9b workflows
  • Krea 2 workflows
  • Windows, macOS or Linux with Python, to download the files
Start with
pip install -U huggingface_hub
hf download Bhavu02/BFS-Best-Face-Swap

Platform

WindowsOperating system
macOSOperating system
LinuxOperating system

About

What it is

BFS, short for Best Face Swap, is a series of LoRA add-ons for image-editing models. They swap a face, a whole head or a full person in a picture, and the author aims for natural skin tones and even lighting. Each release targets one base model: Qwen Image Edit, Flux 2 Klein or Krea 2.

Who it suits

It suits image editors and artists who already run an image-editing workflow and want to replace a face or head in a picture. Use only faces and bodies you have the right to use.

Who uses it
IndustryWho uses itHow they use itExample
Photo editingRetouchersReplace one face in a photo with anotherSwap only the face and keep the target's hair and background
Creative studiosDigital artistsMove a whole head onto a different bodyBlend head, hair and body into one natural picture
Content productionWorkflow buildersReplace the person in a scene with a reference personChange the background through the extra prompt line

People who need to make images from text alone, or who have none of the three base models, can skip it.

Where to use it

  • Qwen Image Edit 2509 and 2511 workflows
  • Flux 2 Klein 4b and 9b workflows
  • Krea 2 workflows
  • Windows, macOS or Linux with Python, to download the files
Top jobs
Use caseWhoWhat they get back
Face-only swapEditorsA new face, with the target's hair, lighting and background kept
Head swapArtistsA full head transferred and blended into the target body
Body swapWorkflow buildersThe person replaced by the reference person, though the pose may differ

How to set it up

  • Python on your computer
  • The huggingface_hub package
  • The base model that matches the LoRA you choose
  • A workflow for that base model
  1. Install the Hugging Face hub package.
  2. Download the model files to your computer.
  3. Pick the LoRA that matches your base model. Versions marked Recommended on the card are the usual choice.
  4. Open the matching workflow. The Krea 2 ones are in the workflows folder of the repository.
pip install -U huggingface_hub
hf download Bhavu02/BFS-Best-Face-Swap

How to use it

Load a LoRA with its own base model and send it two images. For Head V3 and every later version, the body goes first and the face second. Earlier versions take the face first.

Main versions
ToolWhat it does
Face V1Swaps only the face and keeps the target's hair, lighting and background
Head V1 to V2Full head swap, with V2 improving skin tone, pose and anatomy
Head V3 to V5More stable and accurate, with V4 and V5 stronger on expressions
Flux 2 Klein Head V1Head swap for the 4b and 9b Flux 2 Klein models
Krea 2 Head Swap V1.1Updated head swap that replaces the earlier Krea 2 file
Krea 2 Body Swap V1Experimental swap of the whole person

The Krea 2 releases work from a trigger prompt. For a head swap, use this line.

head_swap: replace the head with the reference head.

For a body swap, use this line instead.

body_swap: replace the person with the reference person.

To change the background too, add this line under the trigger.

BG: [describe the new background]

Example

With the Krea 2 head swap LoRA, you give the model a target picture and a reference head, then use the head_swap prompt above. You get the target picture back with its head replaced by the reference head. The card suggests pairing the body swap LoRA with the head swap LoRA at 0.5 strength each to help head identity.

For newcomers and experts

Think of each LoRA file as a small add-on that teaches a large image model one trick. As a first try, take the version the card marks Recommended for your base model and run its workflow on two pictures you own. Mind the order of the two images, because the newer versions expect the body first.

If you know the field, note how Head V5 for Qwen Image Edit 2511 is built. It has an original file trained for more than 5,500 steps, plus merged versions that fold in the V4 head swap. The author's own evaluation says the merged version reproduces a wider range of expressions, though it is not perfect. The Flux 2 Klein 9b files come at rank 128 and rank 64.

  1. Pick the LoRA for your base model
  2. Open the matching workflow
  3. Send the two images in the right order
  4. Get the swapped picture back

Good to know

  • Match the base: use each LoRA only with the base model it was made for.
  • Image order: Head V3 and later take the body first, then the face.
  • Experimental: the body swap may not copy the pose exactly.
  • License: MIT.
  • Cost to run: none named for the LoRA files, and you need the matching base model.
  • Data it reaches: the images you feed it, and downloads from Hugging Face.
  • Not for: swapping in faces or bodies you have no right to use.

README and files

README

BFS Best Face Swap

LoRA add-ons that swap a face, a head or a whole person in an image, built for Qwen Image Edit, Flux 2 Klein and Krea 2. For image editors who have the right to use the faces involved.

Made by Bhavu02 on Hugging Face. These notes are Mindket's, written from its model page and model card; the model's own files and license are on Hugging Face.

About

What it is: BFS, short for Best Face Swap, is a series of LoRA add-ons for image-editing models. They swap a face, a whole head or a full person in a picture, and the author aims for natural skin tones and even lighting. Each release targets one base model: Qwen Image Edit, Flux 2 Klein or Krea 2.

Who it suits: It suits image editors and artists who already run an image-editing workflow and want to replace a face or head in a picture. Use only faces and bodies you have the right to use.

Title: Who uses it
Industry | Who uses it | How they use it | Example
Photo editing | Retouchers | Replace one face in a photo with another | Swap only the face and keep the target's hair and background
Creative studios | Digital artists | Move a whole head onto a different body | Blend head, hair and body into one natural picture
Content production | Workflow builders | Replace the person in a scene with a reference person | Change the background through the extra prompt line

People who need to make images from text alone, or who have none of the three base models, can skip it.

Where to use it:

  • Qwen Image Edit 2509 and 2511 workflows
  • Flux 2 Klein 4b and 9b workflows
  • Krea 2 workflows
  • Windows, macOS or Linux with Python, to download the files
Title: Top jobs
Use case | Who | What they get back
Face-only swap | Editors | A new face, with the target's hair, lighting and background kept
Head swap | Artists | A full head transferred and blended into the target body
Body swap | Workflow builders | The person replaced by the reference person, though the pose may differ

How to set it up:

  • Python on your computer
  • The huggingface_hub package
  • The base model that matches the LoRA you choose
  • A workflow for that base model
  1. Install the Hugging Face hub package.
  2. Download the model files to your computer.
  3. Pick the LoRA that matches your base model. Versions marked Recommended on the card are the usual choice.
  4. Open the matching workflow. The Krea 2 ones are in the workflows folder of the repository.
pip install -U huggingface_hub
hf download Bhavu02/BFS-Best-Face-Swap

How to use it: Load a LoRA with its own base model and send it two images. For Head V3 and every later version, the body goes first and the face second. Earlier versions take the face first.

Title: Main versions
Tool | What it does
Face V1 | Swaps only the face and keeps the target's hair, lighting and background
Head V1 to V2 | Full head swap, with V2 improving skin tone, pose and anatomy
Head V3 to V5 | More stable and accurate, with V4 and V5 stronger on expressions
Flux 2 Klein Head V1 | Head swap for the 4b and 9b Flux 2 Klein models
Krea 2 Head Swap V1.1 | Updated head swap that replaces the earlier Krea 2 file
Krea 2 Body Swap V1 | Experimental swap of the whole person

The Krea 2 releases work from a trigger prompt. For a head swap, use this line.

head_swap: replace the head with the reference head.

For a body swap, use this line instead.

body_swap: replace the person with the reference person.

To change the background too, add this line under the trigger.

BG: [describe the new background]

Example: With the Krea 2 head swap LoRA, you give the model a target picture and a reference head, then use the head_swap prompt above. You get the target picture back with its head replaced by the reference head. The card suggests pairing the body swap LoRA with the head swap LoRA at 0.5 strength each to help head identity.

For newcomers and experts: Think of each LoRA file as a small add-on that teaches a large image model one trick. As a first try, take the version the card marks Recommended for your base model and run its workflow on two pictures you own. Mind the order of the two images, because the newer versions expect the body first.

If you know the field, note how Head V5 for Qwen Image Edit 2511 is built. It has an original file trained for more than 5,500 steps, plus merged versions that fold in the V4 head swap. The author's own evaluation says the merged version reproduces a wider range of expressions, though it is not perfect. The Flux 2 Klein 9b files come at rank 128 and rank 64.

Pick the LoRA for your base model
Open the matching workflow
Send the two images in the right order
Get the swapped picture back

Good to know:

  • Match the base: use each LoRA only with the base model it was made for.
  • Image order: Head V3 and later take the body first, then the face.
  • Experimental: the body swap may not copy the pose exactly.
  • License: MIT.
  • Cost to run: none named for the LoRA files, and you need the matching base model.
  • Data it reaches: the images you feed it, and downloads from Hugging Face.
  • Not for: swapping in faces or bodies you have no right to use.

On Hugging Face, 13 November 2026

  • Task: image to image
  • Library: diffusers
  • Downloads in the last 30 days: 407
  • Likes: 3
  • License: mit
  • Languages: en
  • Based on: black-forest-labs/FLUX.2-klein-9B
  • Tags: diffusers, lora, qwen-image-edit, face-swap, head-swap, body-swap, image-editing, ai-generated, comfyui, flux, klein, krea2

Links

Get the model

Download its files with Hugging Face's command line:

pip install -U huggingface_hub
hf download Bhavu02/BFS-Best-Face-Swap

Then follow the usage in its model card. Nothing in this kit runs by itself.

Files
README.md  6 KB

1 file, 6 KB, downloaded as a zip.

What’s new

  • Version 1 · 2026-11-13

Compare with similar

Listed Hugging Face models in the same category that share the most tags with this one.

ProductDownloads, last 30 daysLikesLicenseLast change
BFS Best Face Swap This page4073mit
Moss-Image-Edit LoRA for ComfyUIShares image-editing · comfyui · lora018gpl-3.0
MOSS-Image-Edit-HQShares image-editing · comfyui · lora090apache-2.0
Qwen-Image-2.1-Next-Scene-LoRAShares image-editing · lora · diffusers4,58171apache-2.0
Qwen-Image-2.1-Multiple-Angles-LoRAShares image-editing · lora · diffusers3,505110apache-2.0
MOSS-Image depth control LoRAShares comfyui · lora · diffusers2,0116apache-2.0
Numbers as Mindket last read them, from 2026-11-12 to 2026-11-13.

Reviews

No ratings yet

Write a review

No reviews yet.

More from Mindket

See all from Mindket

anaksor-ocr

Sokheng · Kit

Reads Khmer and English text from documents, receipts and photos of signs, using a detector for text lines and a recogniser for each line. Suits developers who…

New

Qwen2.5-7B-Instruct as GGUF files in eight quantization sizes, to chat with locally through llama.cpp. For people who want a capable 7.6B model on their own co…

New

ACE-Step 1.5 XL SFT

ACE-Step · Kit

Generates music from text with a 4B diffusion decoder, 50 steps and guidance control, aiming for higher audio quality than the base. For creators with a GPU of…

New

Similar in Productivity

See all in Productivity

Guizang PPT Skill

op7418 · Skill

Skill that makes single-file HTML slide decks, slide illustrations and platform covers in two visual styles, with a presenter mode for rehearsal. For speakers…

New29.8k

Frontend Slides

zarazhangrui · Skill

Builds HTML presentations from scratch or from a PowerPoint file and shows visual style previews to pick from, for people who want web slides with no CSS or Ja…

New32.3k

IPTV

iptv-org · Kit

Collects publicly available live TV channels from many countries into playlists that any video player with live streaming can open. Made for viewers and for de…

New144k