Ubuntu 26.10 will feature Myna, an AI-powered private & local speech-to-text solution
Canonical has announced Myna, a new speech-to-text dictation initiative for Ubuntu desktop that prioritizes privacy and local processing. Designed to feel seamlessly integrated, Myna ensures all voice input stays on your device, aligning with privacy-conscious workflows.
For the initial Ubuntu 26.10 release, Canonical is focusing on the core user experience: users can press a keyboard shortcut, dictate naturally, and see text appear directly in their applications, with clear visual cues indicating when dictation is active. This approach highlights reliability over added complexity in the first iteration.
These privacy controls extend beneath the surface. All speech recognition runs on local hardware without requiring an internet connection after installation. Myna only accesses the microphone when dictation is enabled, processes audio strictly in memory, and never uploads audio recordings to external services.
While the first version targets Ubuntu desktop environments running Wayland and validates support on GNOME, the architectural foundations will allow inclusion of other Linux desktop environments as Myna evolves. Canonical stresses that features such as voice assistants, desktop control, translation, automatic language detection, and extended voice commands will be added only after establishing reliable base functionality.
Myna’s modular design will enable the platform to gain new capabilities in future updates.
- Linux Distro
- Free
- Open Source
Related news
External links
- Introducing Myna: Speech to Text for Ubuntu DesktopUbuntu • Official source
- Myna Announced As Speech-To-Text Solution For The Ubuntu DesktopPhoronix
- Canonical unveils Myna speech-to-text app for Ubuntu 26.10The Register
- Canonical reveal Myna, a speech to text system for Ubuntu LinuxGamingOnLinux
- Canonical's New AI Tool Wants You to Talk to Ubuntu Instead of TypeIt's FOSS
- Ubuntu 26.10 Steps into AI with Local Speech-to-TextLinuxiac
Comments
Some people are okay with AI and it may be beneficial for them, but it should be an opt-in plugin of some sort. Choice should be considered a necessity and remain as such for those who migrate to Linux environments.
Good point but I wished that we would migrate to better working distros like fedora or arch heck even Mint and ZorinOS are getting more users as I’m typing this comment
Don't confuse AI as the technology field, and AI as a marketing stunt.
All commercial voice recognizer tools since 2000s use AI as of machine learning, and neural network more recently, because it's the only way to get, at worse, decent results given how diverse voices, languages and microphone qualities are.
And many daily tools already make use of machine learning and neural network, like most microphones (to remove external noises and improve voice, and you cannot disabling it) and most cameras (to guess the best sensor parameters and to improve overall quality, and you cannot disabling it).
Machine learning is just some advanced technic to guess the best output from a large corpus of examples, and is the basis of LLM, but is not supposed to be smart in any way or take over the world.
Of course it's marketing stunt. AI is being thrown in our phases all the while the orgs try to convince us that it will fulfill their dreams of a better world. All this and AI is still in it's infancy and machine learning has a long way to go to provide good clean data.
Who’s asking for more AI if Canonical cared they could optimize their distribution and let us disable their artificial intelligence for good