Canonical launches Myna speech-to-text preview for Ubuntu Desktop

Brief written by the FOSSWire AI editor from Planet Ubuntu How we write

October 12, 2026 1 min read 13 reads

Key Takeaway

Canonical has released Myna, a local-only speech-to-text tool for Ubuntu, as a technology preview installable via PPA on Ubuntu 24.04 LTS, 26.04 LTS, and 26.10.

Summary

Canonical's Myna project, first announced earlier this year, is now available as a technology preview for Ubuntu Desktop users. The tool provides private, offline speech-to-text dictation that runs entirely on the local machine without uploading audio. It supports multiple recognition engines including Parakeet, Whisper, and FunASR, with optional GPU acceleration on supported NVIDIA hardware. The preview is limited to GNOME on Wayland on amd64/amd64v3 systems.

What Myna offers and its current limits

Local-only processing: No network required after model download; audio never leaves the machine
Engine support: Parakeet, Whisper, and FunASR with varying language coverage, accuracy, and hardware demands
Platform constraints: GNOME on Wayland only; amd64 and amd64v3 architectures
GPU acceleration: Optional NVIDIA GPU support; additional accelerators planned
Scope: Dictation-only tool, not a voice assistant; no desktop control via voice commands
Distribution: Available via PPA (ppa:canonical-desktop-team/myna) with myna-config package

Why it matters

Myna addresses a gap in privacy-respecting accessibility tools on Linux desktops, where most speech-to-text options rely on cloud services. Its open development model and support for multiple engines could attract contributors working on local AI tooling for Linux.

Original reference: ubuntu.com →