Deep Face Cam

Local face swap for desktop workflows.

An open-source macOS and Windows app with a native desktop shell, bundled Python backend, explicit model downloads, and a simpler workflow for image, video, and live camera swaps.

Get supporter builds View face swap examples View source
  • Runs locallyYour data stays on your device
  • Open sourceTransparent and auditable
  • Cross-platformmacOS 12+ and Windows 10/11
Deep Face Cam desktop interface with source and target images selected

Face swap video examples

Simple local workflow

01

Add source & target

Import two photos or frames and keep the original files on your machine.

02

Tune & customize

Adjust alignment, blending, face enhancement, and safety settings.

03

Preview locally

Generate a local preview before you export or switch to live mode.

04

Export or go live

Render media files or use the live camera mode for desktop workflows.

Get supporter builds

Source code is public. Ready-to-run installers are offered as supporter downloads to fund packaging, notarization, testing, and maintenance.

macOS

Apple Silicon & Intel

DMG builds with bundled Python sidecar and ffmpeg tools.

Download for macOS Universal .dmg · Notarization ready
Windows

CPU, DirectML & CUDA

Windows x64 MSI and setup installers for common acceleration paths.

Download for Windows .msi and .exe · SmartScreen friendly path

Open source. Local by design.

Deep Face Cam is licensed under AGPL-3.0. You can inspect the code, build it yourself, fork it, and verify what runs on your machine.

100% local processing
No uploads. No tracking. Your data never leaves by default.
Explicit model downloads
You choose what model files are installed after checksum verification.
Auditable and transparent
React, Tauri, and Python backend code are available in the source repo.
$ git clone https://github.com/DeepFaceCamLabs/deep-face-cam
$ cd deep-face-cam
$ npm install
$ npm run tauri dev
View on GitHub

Models download only after confirmation

The app does not commit model binaries to the source repository. Required files are downloaded into the user app data directory after an explicit prompt and checksum verification.

Required Face swap and face analysis models
Optional GFPGAN and GPEN enhancement models
Local cache Generated CoreML files stay on your machine

FAQ

Can I build it myself?

Yes. The source repository includes build and packaging notes.

Do media files leave my computer?

The core workflow is local. Network access is used for model downloads and project links.

Why are installers paid?

The code is open. Supporter installers fund signing, packaging, testing, and ongoing maintenance.

Can I use it on real people?

Only with consent and lawful use. Do not use it for deception, impersonation, harassment, or non-consensual intimate imagery.