Faceswap

Local AI face-swapping software trains models for photos and videos, with a graphical interface on Windows, Linux and macOS.

Faceswap replaces faces in photos and videos using deep learning. It's for VFX artists, hobbyists and researchers who want to train face-swapping models on their own material and run them locally.

Face extraction: Detect and extract faces from source photos and videos to create training material.
Custom model training: Train a model using images of both faces, with model choices including Phaze-A and Villain.
Photo and video conversion: Apply a trained model to replace faces in source images, with video conversion tools for working with footage.

Faceswap has a graphical interface as well as command-line tools. A project requires source material for both faces and a training stage before conversion; reusing an existing model can reduce training time compared with starting from scratch.

The software is free and open source under GPL-3.0 and runs on Windows, Linux and macOS. A modern NVIDIA GPU with CUDA support gives the best performance. It also supports many AMD GPUs through ROCm on Linux.