camfx
github.com
- Free plan
- Linux
What camfx is
camfx is free, MIT-licensed Linux middleware that applies live camera effects and sends the result to a V4L2 virtual camera through v4l2loopback and FFmpeg. Listed effects include background blur or replacement, brightness adjustment, skin smoothing, face-centered framing, and experimental eye warping. Effects can be managed from a command-line interface or GTK GUI, chained together, and switched live through D-Bus. MediaPipe Tasks supplies segmentation and landmark models, initialized when needed and cached locally. The stated prerequisites are FFmpeg, the v4l2loopback kernel module, Python 3.12, and a camera exposed as /dev/video*. Frame rate depends on webcam formats, USB bandwidth, and whether machine-learning effects are enabled; some webcams cannot sustain 30 FPS at 1080p. Installation is documented through a Fedora-focused alpha RPM or Python virtual environment. The alpha release requires manual v4l2loopback setup, while packaging automation and split packages remain deferred.
Who it suits
camfx is for Linux users who want camera effects delivered through a virtual webcam output and are comfortable installing its listed dependencies.
Good to know, and what to watch
- Background blur and replacement are included
- Command-line and GTK interfaces are available
- Effects can chain and switch through D-Bus
- MediaPipe models cache locally
- MIT-licensed and free
- Requires manual v4l2loopback configuration in alpha
- Performance varies with hardware and effects
- Some webcams cannot sustain 30 FPS at 1080p
- Requires Python 3.12 and FFmpeg
Plans
Every plan and pricecamfx
Free plan
The verdict
camfx provides configurable live effects for a Linux virtual webcam, with both GUI and command-line control. Installation and performance depend on system setup, especially for machine-learning effects.
What the maker publishes
- Audience
- The project is for Linux users who want to apply camera effects through a virtual webcam output. source
- Computer vision
- MediaPipe Tasks provides segmentation and landmark models, which are initialized lazily and cached locally. source
- Controls
- The project includes a command-line interface and a GTK GUI for managing effects. source
- Dependencies
- The README lists FFmpeg, the v4l2loopback kernel module, Python 3.12, and a camera exposed as /dev/video* as prerequisites. source
- Distribution
- The project offers a Fedora-focused alpha RPM release, and its README also documents installation from a Python virtual environment. source
- Effect control
- Effects can be chained and switched live through D-Bus. source
- Effects
- Listed effects include background blur and replacement, brightness, skin smoothing, face-centered framing, and experimental eye warping. source
- License
- The repository provides the software under the MIT License. source
- Output
- It outputs to a V4L2 virtual camera using v4l2loopback and FFmpeg. source
- Performance
- Achievable frame rate depends on webcam formats, USB bandwidth, and whether machine-learning effects are enabled; some webcams cannot sustain 30 FPS at 1080p. source
- Purpose
- camfx is virtual webcam middleware for Linux that adds live effects to camera output. source
- Release limitation
- The alpha release requires users to configure v4l2loopback manually, and the release notes say packaging automation and split packages are deferred. source
Checked against the maker's pages on 9 Oct 2026.