A lightweight, native macOS global dictation tool built in Swift and SwiftUI with zero external dependencies.
WhisperMic allows you to record audio from your microphone using a global keyboard shortcut, transcribe it using OpenAI's Whisper API, and automatically paste the transcription into any active text field. It serves as a free, open-source, and privacy-friendly alternative to paid dictation utilities.
- Global Hotkey Activation: Press
Option + Space(configurable) to start recording from anywhere. - Animated HUD: Displays a translucent, glassmorphic floating HUD overlay showing real-time recording waveforms and timers.
- Optimized Audio: Records mono, 16kHz audio in compressed AAC (
.m4a) format to minimize API upload latency. - Secure Storage: Your OpenAI API Key is stored safely in the native macOS System Keychain.
- Universal Paste Injection: Simulates keystroke pasting (
Cmd + V) to automatically insert transcribed text into any application without requiring complex accessibility setups.
WhisperMic is compiled natively from source using standard macOS Command Line Tools.
- macOS 14.0 (Sonoma) or newer.
- Xcode Command Line Tools installed (run
xcode-select --installin Terminal if missing). - A valid OpenAI API Key.
- Open Terminal in the root folder.
- Run the build script:
./build.sh
- Open the generated
WhisperMic.appapplication bundle.
- On first launch, the Preferences panel will open automatically.
- Enter your OpenAI API Key and click Save Key (you can click Test Connection to verify).
- To enable the paste feature, click Grant Access to open System Settings, and check the box next to WhisperMic under Privacy & Security -> Accessibility.
This project is open-source. Feel free to use, modify, and distribute it.