Voice to text, right on your screen Beta

Hold a hotkey, speak, and the text appears automatically at your cursor position. Fully offline — no cloud, no API keys, no data leaves your machine.

AutoWhisper is currently in beta. The app is fully functional, but we're actively working on improvements and new features. Feedback is welcome via GitHub.

Download free View on GitHub

Available for Windows 10/11 only

AutoWhisper

100%

Offline processing

28

Supported languages

GPU

Accelerated (CUDA & Vulkan)

MIT

Open source licence

How does it work?

From voice to text in three simple steps.

Step 1

Hold the hotkey

Press Ctrl+Shift+Space (or your custom hotkey) to start recording. A visual overlay appears at the top of your screen.

Step 2

Speak

Speak naturally into your microphone. AutoWhisper records as long as you hold the hotkey.

Step 3

Text appears automatically

Release the hotkey and the transcribed text is automatically pasted at your cursor position — in any application.

Why AutoWhisper?

Complete privacy

All processing happens locally using OpenAI's Whisper model. Audio is processed in-memory only — no files are saved and no data leaves your machine.

GPU acceleration

Auto-detects CUDA (NVIDIA) and Vulkan (AMD/Intel/NVIDIA) with CPU fallback. Fast transcription regardless of your hardware.

28 languages

Automatic language detection or manual selection. Support for 28 languages including Dutch, English, French, German and more.

Flexible models

Choose from 5 Whisper models, from Tiny (39 MB) to Large v3 (3.1 GB). Download directly from the app with progress tracking.

Runs in the background

Runs silently in the system tray. Optional launch at Windows startup. Minimal system resource usage when idle.

Works everywhere

Compatible with any application that accepts text input — editors, browsers, chat apps, email clients and more.

System requirements

Minimum

  • Windows 10 / 11
  • Modern x64 processor
  • 4 GB RAM
  • 200 MB disk space (+ model size)

Recommended

  • NVIDIA GPU with CUDA support
  • AMD/Intel GPU with Vulkan support
  • 8 GB RAM or more
  • 4 GB disk space (for Large v3 model)

Ready to type faster?

Download AutoWhisper for free and experience voice-to-text without compromising on privacy.

Download for Windows View source code on GitHub

Available for Windows 10/11 only