Private speech-to-text for macOS and iOS. Powered by OpenAI Whisper running entirely on-device with Metal GPU acceleration. No cloud. No subscription. No data leaves your machine.
A menu bar app that stays out of your way until you need it.
Everything runs on your hardware. Models download once, then work entirely offline.
Text appears as you speak with word-level stability detection. No waiting for silence — your words flow directly into any text field.
Record entire meetings with automatic speaker diarization. Each speaker is identified and labeled in the final transcript.
whisper.cpp runs natively on Apple Silicon via Metal, delivering fast inference without sending audio to any server.
Zero network requests for transcription. Your audio is processed locally and never recorded, uploaded, or stored by third parties.
Optionally refine meeting transcripts with OpenAI, Anthropic, or Claude Code CLI. Identify speakers by name and fix errors.
Interface available in English, French, Spanish, Chinese, Portuguese, and German. Whisper transcribes 99 spoken languages.
The menu bar icon displays the active language as a simplified flag pattern, giving instant visual feedback without opening Settings.
Assign a keyboard shortcut to cycle through your preferred languages instantly while typing.
Switch input devices directly from the menu bar without opening Settings. Perfect for toggling between headset and built-in mic.
Found a bug? Settings → Support opens a pre-filled email to the developer including your description and recent app logs.
Three steps, no accounts, no cloud dependencies.
AVAudioEngine records at 16 kHz directly from your microphone. On macOS, hold a global hotkey; on iOS, tap the record button.
whisper.cpp processes audio entirely on-device using Metal GPU. Choose from base, small, medium, or large models depending on your accuracy/speed needs.
On macOS, text is injected directly at your cursor position in any app via CGEvent. On iOS, text is copied to clipboard or saved to your history.
MacWhisperer for your desktop, iWhisperer for your pocket. Same engine, tailored for each platform.
| MacWhisperer | iWhisperer | |
|---|---|---|
| Menu bar app | ✓ | — |
| Global hotkey (hold to speak) | ✓ | — |
| Text injection at cursor | ✓ | — |
| Streaming transcription | ✓ | — |
| Meeting recording + diarization | ✓ | — |
| LLM transcript refinement | ✓ | — |
| Transcription history (SwiftData) | — | ✓ |
| Live Activity + Dynamic Island | — | ✓ |
| Siri Shortcuts / Action Button | — | ✓ |
| Onboarding flow | ✓ | ✓ |
| Metal GPU acceleration | ✓ | ✓ |
Clone the repo, build whisper.cpp, and run. That's it.