Launching today
Phantom Voice
Local push-to-talk dictation for Mac
36 followers
Local push-to-talk dictation for Mac
36 followers
Push-to-talk dictation that inserts in place: hold fn, speak, release, and the text appears in the app your cursor is already in. Transcription runs locally with whisper.cpp + Metal, so your audio never leaves the Mac. An optional on-device cleanup pass strips filler words and fixes punctuation. Most dictation apps show you a recording dot or a generic soundwave. This one has hand-animated indicators. $10 once, no subscription. MIT source. macOS 14+, Apple silicon.
Products used by Phantom Voice
Explore the tech stack and tools that power Phantom Voice. See what products Phantom Voice uses for development, design, marketing, analytics, and more.
Whisper by OpenAIA neural net for speech recognition
5.0 (38 reviews)
Phantom Voice's whole premise is that your audio never leaves the Mac, so a cloud endpoint was disqualifying regardless of accuracy — the moment audio goes over the wire there's nothing left to differentiate on.
Whisper is the only one here with open weights, which is what makes on-device inference possible at all. Specifically I ship whisper.cpp, Georgi Gerganov's C/C++ port, compiled with Metal so it runs on the GPU and the model stays resident between dictations rather than reloading per utterance.
It also costs nothing per transcription, and that's what lets the app be $10 once instead of a subscription. A per-minute API would have forced recurring pricing just to cover inference — so this choice determined the business model, not only the architecture.