VoxInput

VoxInput is an open-source dictation tool for Linux and macOS that types spoken words into the active application. It's for people who want speech input across their desktop, including applications without their own dictation feature. On Linux, it also works in the system console.

The app runs on your machine, while a local or remote service handles transcription. It's designed for LocalAI and works with OpenAI-compatible APIs that provide transcription or realtime audio. With a local backend, audio processing can stay on your own hardware. Choosing a remote service sends audio to that service instead. A local instance doesn't have to require an API key.

VoxInput can capture your microphone or system audio, so it also serves as a transcription tool for meetings and video streams. It can type the resulting text into an application or save it to a file. Realtime transcription detects speech automatically; a separate recording mode transcribes after recording stops.

Noise suppression reduces background sound in microphone recordings. LocalVQE provides audio cleanup, with compact models suited to low-power hardware such as Raspberry Pi 5. Recording notifications show when speech capture starts and stops, and a terminal interface provides recording controls alongside chat and log views.

VoxInput uses dotool for text entry on Linux and CoreGraphics on macOS. The macOS app needs Accessibility permission to simulate input. The project uses the MIT license.