Faceswap
Free, open-source deepfake software for training your own face-swap models
- Category
- Video Generation
- Pricing
- Completely free and open source under GPL-3.0; project is Patreon and donation supported
- Best for
- Best suited to developers, AI/ML hobbyists, and VFX practitioners who are comfortable with Python and command-line tools and want a completely free, locally-run deepfake pipeline with full control over training data and models.
- Official site
- faceswap.dev
- Last updated
- August 2026
Faceswap (faceswap.dev) is a free, open-source deepfake software project that uses deep learning to recognize and swap faces in photos and videos. Built on TensorFlow, Keras, and Python, it runs locally on Windows, macOS, and Linux rather than as a cloud SaaS product. The core workflow follows three stages: Extract pulls faces out of raw source photos or video frames using facial landmark detection; Train uses two folders of face images to train a neural network model, with support for reusing existing models to speed up training; and Convert applies the trained model to swap faces into target images or video. The project also ships a GUI for users who prefer not to work purely from the command line, plus a bundled ffmpeg-based tool for converting video into photo sequences and back.
The project traces back to the original 'deepfakes' code that first made face-swapping AI accessible outside academic research, and it has grown into one of the two dominant open-source deepfake tools alongside DeepFaceLab, with Faceswap generally regarded as more beginner-friendly and better documented, while DeepFaceLab is seen as more powerful for advanced professional results. Faceswap is maintained by a volunteer community rather than a company, with key contributors like torzdf (GUI, aligners, model porting) and andenixa (Unbalanced and OHR models) funded through Patreon and one-time donations. The project publishes an explicit ethical manifesto stating the software is not intended for non-consensual or inappropriate content, and the maintainers say they take a zero-tolerance stance on unethical use while supporting the tool's legitimate applications in VFX, research, and commentary.
Best suited to developers, AI/ML hobbyists, and VFX practitioners who are comfortable with Python and command-line tools and want a completely free, locally-run deepfake pipeline with full control over training data and models. It is a poor fit for non-technical users wanting instant, one-click face swaps, or anyone needing official commercial support and guaranteed turnaround.
Key features
Extract
Detects and crops faces from a folder of source photos or video frames into a dedicated extract folder, ready for training.
Train
Trains a custom neural network model from two folders of source and destination face images, saving progress into a models folder; existing models can be reused to speed up training.
Convert
Applies a trained model to swap faces into a target set of images or video frames, producing the final output.
GUI and CLI
Offers both a graphical interface (python faceswap.py gui) and full command-line control with help flags on every script.
Multiple Model Architectures
Supports several swappable model types including Villain, DFL-H128, DFaker, Phaze-A, Unbalanced, and OHR, each with different quality/speed tradeoffs.
GPU Acceleration
Leverages CUDA-capable GPUs for training speed, with ROCm support extending GPU acceleration to many AMD cards on Linux.
ffmpeg Conversion Tool
Includes a bundled effmpeg utility to convert video into photo sequences and back into video for the extract/convert pipeline.
Community Support Channels
An active Discord server and dedicated forum provide troubleshooting help, guides, and tutorials outside the GitHub issue tracker.
Pricing breakdown
Open Source (Self-Hosted)
- Full source code and all features included
- Runs on your own Windows, macOS, or Linux hardware
- No usage caps, credits, or subscription
- Optional Patreon/one-time donations support core developers
Pros and cons
Pros
- The software and its source code are entirely free under the GPL-3.0 license, with no subscription, credits, or per-generation fees, so cost is limited to hardware and electricity.
- With 57.1k GitHub stars, 13.5k forks, and 1.5k watchers, Faceswap is one of the most established and actively maintained open-source projects in the face-swap space, indicating a large contributor and user base.
- Running entirely on local hardware gives users full control over sensitive source material, since nothing is uploaded to a third-party server for processing.
- Cross-platform support for Windows, macOS, and Linux, plus AMD ROCm compatibility on Linux, means users aren't locked into a single OS or NVIDIA-only GPU ecosystem.
- The Extract/Train/Convert pipeline is well documented via INSTALL.md and USAGE.md, and the project maintains an active Discord and forum specifically for user support and tutorials.
- Choice of multiple model architectures (e.g., Villain, DFaker, Phaze-A) lets advanced users tune the quality-versus-speed tradeoff for their specific hardware and use case.
Cons
- The tool has a genuinely steep learning curve: it requires comfort with command-line workflows, GPU driver setup, and understanding of the extract/train/convert pipeline, which is a barrier for non-technical users.
- Producing convincing results requires a capable GPU and often many hours or days of model training, unlike instant cloud-based face-swap apps.
- There is no official customer support, warranty, or SLA; all troubleshooting happens through community Discord and forums, which can mean slower or inconsistent help.
- As deepfake software, it carries inherent ethical and reputational risk; despite the project's explicit anti-abuse manifesto and zero-tolerance policy, misuse by third parties remains a real-world concern the maintainers must continually address.
- General support questions are explicitly not welcome in the GitHub repo issue tracker and must be redirected to the forum, adding friction for users unfamiliar with the project's norms.
Alternatives to Faceswap
Runway
Generative video and editing suite for creators
Compare →Synthesia
AI avatar videos for training and corporate content
Compare →HeyGen
AI avatar and localization tool for marketing video
Compare →Kling AI
AI video generator known for realistic motion
Compare →Shai
AI pre-production for film and advertising — script to storyboard in seconds
Compare →RepliQ
Personalized video, image, and text outreach at scale
Compare →Luma Dream Machine
Cinematic AI video and image generation from a single prompt
Compare →Rask AI
Dub your videos into 130+ languages in your own voice
Compare →Vyond
Animated and photorealistic AI video without a production crew
Compare →Fliki
Turn scripts and ideas into narrated video in minutes
Compare →Sendspark
Record one video, AI personalizes it for thousands of prospects
Compare →Adori AI
Turn blogs, text, and audio into AI-generated videos in minutes
Compare →Flixier
Cloud-based video editor with AI generation built into the timeline
Compare →Hedra
AI character videos from a single photo and audio clip
Compare →MakeShorts.ai
Turn long YouTube videos into viral shorts instantly
Compare →MiniMax Video (Hailuo AI)
Cinematic AI video generation with native audio, budget pricing
Compare →Neural Frames
Turn songs into AI music videos with audio-reactive visuals
Compare →Neural.love
All-in-one AI studio for video, image, and audio creation
Compare →Plotagon
Turn scripted text into acted-out animated movies
Compare →Synthesys AI Studio
AI avatars, voices, and UGC video ads in one plan
Compare →TopMediai
All-in-one AI studio for video, music, and voice
Compare →VideoScribe
Whiteboard-style animated explainer videos, no design skills needed
Compare →Wisecut
AI video editor that auto-cuts, captions, and reframes clips
Compare →Frequently asked questions
Is Faceswap really free?
Yes, it is fully free and open source under the GPL-3.0 license; the developers accept optional Patreon and one-time donations but the software itself carries no cost or paywalled features.
Do I need a powerful GPU?
A modern CUDA-capable GPU (or an AMD GPU with ROCm support on Linux) is strongly recommended for practical training times; the project's INSTALL.md details minimum requirements.
How is Faceswap different from DeepFaceLab?
Both are open-source deepfake tools, but Faceswap is generally considered more beginner-friendly with wider documentation and a larger contributor base, while DeepFaceLab is viewed as offering more advanced control at the cost of a steeper learning curve.
What operating systems are supported?
Faceswap runs on Windows, macOS, and Linux, powered by TensorFlow, Keras, and Python, with GUI and command-line options on all three.
Is Faceswap legal and ethical to use?
The project explicitly states it is not for creating non-consensual or inappropriate content and maintains a zero-tolerance policy on unethical use, positioning itself for research, VFX, commentary, and other legitimate applications; users remain responsible for complying with local laws.
Where do I get help if something breaks?
Support is community-driven through the official FaceSwap Discord server and forum; general support questions are not handled in the GitHub issues tracker, which is reserved for code-level bugs.
Ready to try Faceswap?
Head to the official site to explore pricing and start a free trial where available.
Visit Faceswap →