How live face swap works: three steps to live
No dataset to gather, no model to build, no export queue. Load your photo and voice clip, tune the look, and go live with a real-time face swap on any video call, in any app, with one click.
Pick your source input
Select a source photo for the face swap; several photos of one person give a more stable identity, and you can pick another photo at any time. Add a short audio clip (a few seconds is enough) as your real-time voice reference.
Dial in the look and sound
Choose your swap engine and resolution, enhance detail, restore expressions, add anime or cinematic looks, reshape or re-age the face, tune your cloned voice, then pick your camera and mic.
Go live
Hit Go Live and NZCAM streams your swapped face and cloned (or translated) voice into the NZCAM virtual camera and virtual microphone; pick them in WhatsApp, Telegram, Zoom, Teams, Discord, Twitch, Facebook, Instagram, YouTube, Line, WeChat, Google Meet, etc., exactly like a normal webcam and mic. Hit Stop to end the session.
Three modes
Pick the mode that fits how you're using NZCAM.
Live
Webcam → virtual camera and microphone, for streaming and calls.
Image / Video
Process a photo or video file to a file; the full realism suite and colour looks apply.
Headless / CLI
Run NZCAM without the UI for scripted or automated workflows.