AirType
Turn your voice into text anywhere on your computer - offline, private, instant.
The Pitch
Press a hotkey. Speak. Your words appear wherever your cursor is - in any app, on any screen. Run it fully offline with local Whisper AI, or connect your OpenAI API key for premium streaming accuracy. Your choice, your data.
How It Works
1. Press Your Hotkey
Configure any hotkey you want - works anywhere on your computer
2. Speak
Talk naturally in English or Hebrew
3. Text Appears
Release and your words type themselves
Key Features
100% Offline Option
Local Whisper AI runs entirely on your machine. Your voice never touches the cloud - perfect for sensitive work, legal docs, medical notes, or just privacy-conscious users.
Multilingual
Dedicated hotkeys for English and Hebrew (with RTL support). Switch languages mid-workflow without touching settings.
Lightweight
Built with Tauri + Rust, not Electron. The entire app is under 10MB (vs 150MB+ for typical desktop apps). Near-zero CPU when idle.
Works Anywhere
Global hotkeys mean it works in any application - browsers, terminals, Word, Slack, email clients, code editors. Wherever you can type, you can speak.
Two Modes
FREE - Local Whisper
- • Completely offline
- • Download model once (75MB - 3GB)
- • Great accuracy
- • No API costs ever
PREMIUM - OpenAI Realtime
- • Live streaming transcription
- • Best-in-class accuracy
- • Bring your own API key
- • Pay per use
Tech Under The Hood
- ▸ Tauri v2 - Native desktop framework (Rust backend, web frontend)
- ▸ whisper-rs - Rust bindings to whisper.cpp for local ML inference
- ▸ Svelte 5 - Reactive UI for settings and status display
- ▸ enigo - Cross-platform keyboard simulation for text injection
- ▸ WebSocket - OpenAI Realtime API for premium streaming mode
Why I built this: I wanted voice typing that worked everywhere without sending my audio to the cloud. Existing solutions were either web-only, required subscriptions, or were bloated Electron apps. AirType is what I wanted: fast, private, and it just works.