Canonical introduces Project Myna, bringing private, locally-processed voice dictation natively to Ubuntu 26.10 desktops.
Canonical has announced Project Myna, a new initiative aimed at integrating native speech-to-text dictation directly into the Ubuntu Desktop ecosystem. Named after the vocal mimicry of the myna bird, the project aims to make voice typing a first-class, highly productive accessibility feature for Linux users.
Slated to debut with the Ubuntu 26.10 release, Project Myna is intentionally narrowing its initial scope to master fundamental desktop dictation. The user workflow is simple: users press a designated keyboard shortcut, speak naturally, and see the resulting text seamlessly appear inside their active application. To ensure a stable rollout, the initial release targets Ubuntu Desktop running on Wayland, with GNOME serving as the primary validated desktop environment. Advanced AI features like voice assistants, custom voice commands, desktop control, and language translation are explicitly sidelined for future releases.
The framework prioritizes speech recognition models that run entirely locally on the user’s hardware. Once the required local models are installed, an active internet connection is not required to use the service. Furthermore, microphone permissions are strictly gated—activating only upon user invocation—and audio data is processed in volatile memory and instantly discarded rather than being shipped to external servers.
Canonical emphasizes that Myna is a highly modular platform built to evolve. The development team is actively seeking early community feedback from testers, developers, and assistive technology users to shape the system before its official implementation.
