Effective: March 23, 2026 · Updated October 7, 2026
Each Faraday Soft app has its own policy. This page covers Voice for Mac. For Lock Chimes for iPhone, see the Lock Chimes privacy policy.
Voice is made by Faraday Soft, the app studio of Leon Johnson, in California, USA. Contact: support@faradaysoft.com
Nothing. Voice does not collect, store, or transmit any data from your device. Specifically:
Voice uses whisper.cpp, an open-source speech recognition engine that runs entirely on your Mac's CPU. Your audio is recorded, transcribed, and deleted locally. At no point does audio data leave your machine.
Voice can optionally clean up text using a local llama.cpp backend and a model file stored on your Mac. This feature is:
Voice is free and open source (GPL-3.0). There is no purchase, no account, no license key, and no payment processor involved. Nothing about you is collected or transmitted for billing or activation.
Don't take our word for it. You can inspect the app's network activity:
sudo lsof -i -n -P | grep Voice
During normal dictation you should not see remote endpoints. Transcription and cleanup run in-process on your Mac. The only network requests Voice makes are one-time model downloads from Hugging Face (verified against pinned checksums) and, if you use it, the "Check for Updates" menu item.
If we ever change this policy, we will update this page. Our commitment to local audio and transcription processing is fundamental to Voice and will not change.