Professional desktop voice changer
Run our native app on Windows with NVIDIA GPU acceleration, route through a VB-Audio virtual cable, and use your transformed voice as the microphone in Teams, Zoom, Google Meet, Discord, Messenger, and any app that accepts a mic input.
Installer link coming soon — try the live web app
Open live voice changerBuilt for NVIDIA GPUs
The desktop stack targets CUDA-capable NVIDIA hardware on Windows for fast inference. Keep your GPU drivers up to date for the best real-time experience.
Virtual microphone for every chat app
Voicechanger.co outputs processed audio to a virtual playback device (we recommend VB-Cable). In Zoom, Microsoft Teams, Google Meet, Discord, or Messenger, choose the cable’s microphone side as your input so callers hear your AI-transformed voice — not your raw mic.
Works where you already talk
Microsoft Teams · Zoom · Google Meet · Discord · Facebook Messenger · Skype · Slack calls — any software that lets you pick an input device.
VB-Audio VB-Cable (recommended)
Install the free VB-Cable virtual audio device from VB-Audio. In our app, send output to CABLE Input (VB-Audio Virtual Cable). In your meeting app, select CABLE Output as the microphone. VB-Audio is third-party software; follow their installer and docs.
Get VB-Cable at vb-audio.comSignal flow in one glance
- 1Your headset or USB mic → Voicechanger.co desktop (capture)
- 2GPU-accelerated AI transforms speech in near real time
- 3Processed audio → VB-Cable CABLE Input (playback device)
- 4Meeting app mic → CABLE Output (so others hear the changed voice)
Windows 11 audio tips
- Prefer DirectSound for capture when the app exposes host APIs — it avoids common WDM-KS quirks with some virtual cables.
- Confirm Windows Privacy → Microphone allows desktop apps to access the mic.
- Use headphones to reduce echo when monitoring your own voice.
Low latency, phrase by phrase
Live mode segments speech into phrases using voice-activity detection, runs ASR and cloned-voice synthesis on the GPU, then plays each phrase to your virtual cable. Tune behavior with desktop environment variables such as OMNIVOICE_LIVE_UTTERANCE, OMNIVOICE_LIVE_END_SILENCE_MS, and OMNIVOICE_LIVE_CHUNK_SECONDS (see desktop documentation).