your voice, instantly typed.
press-and-hold: hold hotkey while speaking, release to transcribe
lock mode: double-tap hotkey to start recording, single tap to stop and transcribe
built with the swift composable architecture and powered by whisperkit.
app/- main app entry point and lifecycle managementclients/- core services (recording, transcription, pasteboard, hotkeys)features/- app features (transcription, settings, history)models/- data models and settingsresources/- assets, sounds, and language data
- apple silicon mac (m1 or later)
- macos 13+ (sonoma recommended)
- xcode 15.x (swift 5.9)
important: don't use xcode 16/swift 6.0 yet - macro compatibility issues.
-
install xcode 15
- download from apple developer
- extract and move to
/applications
-
set xcode 15 as active
sudo xcode-select -s /Applications/Xcode_15.app
-
clone and build
git clone <repo-url> cd vocorize xcodebuild -resolvePackageDependencies -project Vocorize.xcodeproj -scheme Vocorize
-
run in xcode
- open
vocorize.xcodeproj - select
vocorizescheme - press
⌘rto run
- open
- macro errors? make sure you're using xcode 15.x
- permission issues? grant mic and accessibility access on first run
- switch back to xcode 16?
sudo xcode-select -s /Applications/Xcode.app
- download and install
- grant microphone and accessibility permissions
- on first run, select the recommended model and download it
- set your hotkey in settings
- start talking
- platform: macos (apple silicon only)
- permissions: microphone + accessibility (to paste text)
- engine: whisperkit (runs locally)
- architecture: swift composable architecture
- offline: no internet required, no data collection
questions? ideas? reach out on twitter.
MIT License - see LICENSE file for details.