Earlier this month plans were shared publicly of Ubuntu 26.10 aiming to build a context-aware desktop with local AI features and one of the first capabilities to be integrated speech-to-text support. Now we have more details on the speech-to-text plans with Canonical announcing the Myna project.

Myna is the speech-to-text solution being developed by Canonical for the Ubuntu desktop. So far in the GPLv3-licensed GitHub repository for Myna is just the documentation and is described as:

"The project draws its name from the myna, a bird renowned for its ability to listen to, mimic, and reproduce human speech with astonishing clarity. Just like its avian counterpart, this application is designed to master voice audio-listening intently to your spoken words and instantly translating them into accurate, clean text. Whether you are looking to dictate text hands-free, improve accessibility, or streamline your workflow, myna brings seamless voice recognition directly to your Linux ecosystem."

Jean Baptiste Lallement of the Canonical Desktop Team shared more details on Myna via the Ubuntu Discourse. For Ubuntu 26.10 they are aiming for the "basics" with reliable desktop dictation:

"The initial experience will be simple: Press a keyboard shortcut, speak naturally, and see the resulting text appear in the application you’re using. Myna is designed to provide speech recognition with clear visual feedback while dictation is active.

Under the hood, Myna uses speech recognition models running locally on your machine. The initial release targets Ubuntu Desktop on Wayland, with GNOME as the primary validated environment, while keeping the architecture open enough to support additional desktop environments in the future."

For this initial Ubuntu 26.10 milestone they are not aiming for voice assistants, voice commands, desktop control, translation, or other features but for now just the desktop dictation for speech-to-text.

Myna supports all local processing for speech models, thus no Internet connection is required after the language model has been downloaded.

Those wishing to learn more about the Myna project can do so via Ubuntu Discourse.