Privacy-first. Blazing fast. Hold a key, speak, release — text appears where you need it.
Global hotkey captures audio instantly. Release to transcribe. Text injected into your active app.
All transcription runs locally. Audio and logs never leave your device unless you explicitly opt in.
Written in native code with GPU acceleration. The fastest local transcription engine available.
Use your own API key for cloud LLMs. Enable or disable at any time — your choice, always.
Auto-detect, CPU-only, or GPU acceleration. Choose the model size that fits your hardware.
Route audio through Gemini, GPT, or specialized models for ultra-accurate transcription of technical content.
30 seconds of English audio
Every transcription runs on your hardware. Your audio never leaves your machine. Activity logs stay local. Online services are available — but only when you choose, with your own API keys.
Runs entirely on your CPU or GPU
Zero data collection by default
Online features use your own credentials