Ubuntu 26.10 is notably working on [1]laying the foundation for a context-aware desktop and their initial deliverable being worked on is [2]Myna as a speech-to-text solution for the Linux desktop. Interestingly there is already a promising voice dictation / speech-to-text solution for the Linux desktop called Vocalinux that continues advancing and is usable right now for those looking at their own speech-to-text desktop solution.

Vocalinux is a GPLv3-licensed solution that works both on X11 and Wayland desktops under Linux for completely offline speech-to-text handling. Vocalinux leverages Whisper.cpp for the microphone audio processing while alternatively VOSK or OpenAI Whisper with PyTorch can be used too. With Whisper.cpp you can enjoy Vulkan-based acceleration that works across GPU vendors or there is the CPU back-end as well.

Vocalinux supports a desktop tray UI, configurable hotkeys for activation, and other typical features one would expect for interacting with your desktop via voice dictation. Compared to the in-development Canonical Myna, it aims to be Linux distribution agnostic as well as desktop agnostic.

Yesterday saw the release of [3]Vocalinux 0.14 beta that landed the configurable keyboard shortcuts support, FunASR/SenseVoice support for remote API engines if desired, some Wayland fixes, and better handling of the CPU-based Whisper.cpp support on hybrid CPU laptops.

Those wishing to learn more about this open-source Linux desktop voice dictation software can do so via [4]GitHub and the project site at [5]Vocalinux.com .



[1] https://www.phoronix.com/news/Ubuntu-26.10-Desktop-Features

[2] https://www.phoronix.com/news/Ubuntu-Myna-Speech-To-Text

[3] https://github.com/jatinkrmalik/vocalinux/releases/tag/v0.14.0-beta

[4] https://github.com/jatinkrmalik/vocalinux

[5] https://vocalinux.com/