Channel Partner Sales Executive

How to try out 26.10's speech-to-text app in Ubuntu 26.04 | 24.04. Published: September 29, 2026 Discover more Open Source Linux & Unix This is a step by step guide shows how to install and try out Myna, Ubuntu's new offline speech to text feature in Ubuntu 26.04 and Ubuntu 24.04 LTS. Canonical, the company behind Ubuntu, is developing a voice typing application called Myna. It's an opt-in feature in Ubuntu 26.10, while the current Ubuntu 26.04 can also try it out. How it works. Just press a keyboard shortcut and speak naturally, the app will record what you said, then use an AI model to convert the voice to text, finally type the text in your application. All the process works offline! And, myna is made up of three components: * myna-config - the graphical configuration tool, responsible for choosing which AI model to use (Whisper, Parakeet, or FunASR), set keyboard shortcut, language, etc. * myna client and AI models, installed as snap packages. * Gnome Shell extension for dictation activity indicator. Step 1: install myna-config graphical configuration tool. For user on modern Intel/AMD platform, Ubuntu developer maintains a PPA that contains myna-config package for Ubuntu 24.04 and Ubuntu 26.04. Linux & Unix To install the package, first press Ctrl+Alt+T on keyboard to launch terminal, then run command to add PPA: sudo add-apt-repository ppa:charles05/myna-config Next, run command to install the configuration tool: sudo apt install myna-config If you're following this tutorial in Linux Mint, then run sudo apt update to refresh cache first. Finally, launch the tool from either Gnome Overview or by running myna-config in terminal. Linux & Unix Step 2: install myna client and AI models. Both the client app and AI models are available as Snap packages. To install then, first launch terminal (Ctrl+Alt+T) then run the commands below one by one: * First, run command to allow snap packages to run background services as user: sudo snap set system experimental.user-daemons=true * Then, install the myna client (so far available in edge repository): sudo snap install myna -edge * Finally, download & install AI models (choose one that you prefer, or install all of them): sudo snap install -edge myna-parakeet sudo snap install -edge myna-whisper sudo snap install -edge myna-funasr Step 3: install GNOME Shell extension (only for default GNOME desktop). For better integration, you may install myna-shell extension that can display small on-screen overlay showing the status of dictation. The extension so far is not available in extensions.gnome.org, so you need to manually install it from source. * First, launch terminal (Ctrl+Alt+T) and run command to grab the myna source code: git clone https://github.com/canonical/myna.git Run sudo apt install git to install if git does not exist. * Create the extension directory: mkdir -p ~/.local/share/gnome-shell/extensions/[email protected] * Finally, copy the extension from source to the directory you just created: cp -r myna/extensions/myna-shell/* ~/.local/share/gnome-shell/extensions/[email protected] After installed the extension, log out and back in. Then, launch "Extension Manager" (available in App Center / Ubuntu Software as Debian package) and enable that extension. Configure and use myna. Now, launch myna-config, choose AI model, set a keyboard shortcut, your spoken language (e.g., en) and other. Finally, launch a text editor or application to input, press the keyboard shortcut and speak naturally, and see if it works! Uninstall: If you don't like the new voice typing application, or it does not work properly for you, then launch terminal (Ctrl+Alt+T) and remove it by following steps. * First, remove the extension either by using "Extension Manager" or by simply removing the source: rm -R ~/.local/share/gnome-shell/extensions/[email protected] * Then, remove the client and AI models: snap remove -purge myna myna-parakeet myna-funasr myna-whisper * Finally, remove the configuration tool: sudo apt remove myna-config Also remove the PPA afterward by running command: sudo add-apt-repository -remove ppa:charles05/myna-config Enable this blog? Please spread the world:)

Canonical announces the alpha release of Charmed OpenShell to help secure autonomous AI agent fleets. Canonical teams up with NVIDIA on the NVIDIA Open Agent Safety Platform and the Broad Availability of NVIDIA OpenShell, delivering developer-friendly packaging and enterprise-grade infrastructure. Canonical is pleased to announce comprehensive agent-ready enterprise infrastructure for NVIDIA OpenShell which reaches broad availability today, and is part of the new NVIDIA Open Agent Safety Platform. As autonomous, self-evolving AI agents transition from experimental pilots into production operations, NVIDIA OpenShell provides an open source runtime that governs how agents execute, what resources they can access, and where inference traffic is routed. Canonical is teaming with NVIDIA to provide seamless deployment paths ranging from local developer workstations to production-scale enterprise infrastructure. Frictionless desktop experience for developers. The OpenShell snap was released in June 2026 to give developers straightforward access to governed agent runtimes. By packaging OpenShell as a snap, Ubuntu users benefit from a one-command installation, workload confinement, and automated updates. With the OpenShell snap, users can instantly run agent sandboxes locally and test autonomous workflows, without manually configuring complex runtime dependencies. Announcing the alpha release of Charmed OpenShell. To streamline enterprise deployment and lifecycle management, Canonical is announcing the alpha release of Charmed OpenShell. Developed in the open, Charmed OpenShell pairs OpenShell's control plane with Canonical's Juju operator ecosystem, providing a trusted, repeatable way to run and manage the gateway alongside the rest of the platform. Charmed OpenShell deploys and operates the OpenShell Gateway service on Canonical Kubernetes running on MicroCloud, automatically connecting the central control plane into essential enterprise services. This enables: * High availability (HA) and state persistence: automated integration with Charmed PostgreSQL for state store resilience and optimistic concurrency across gateway replicas * Identity and access management: native integration with Canonical Identity Platform for OpenID Connect (OIDC) user authentication and access control * gRPC ingress: managed Transmission Control Protocol (TCP) routing and end-to-end Transport-Layer Security (TLS) passthrough, powered by Traefik * Centralized observability: pre-configured binding to the Canonical Observability Stack (COS), delivering native Prometheus metrics, Grafana dashboards, and Loki log aggregation Enterprise agent infrastructure on MicroCloud. For organizations scaling agent fleets across multiple business units, Canonical provides agent-ready infrastructure through MicroCloud, its lightweight, low-touch cloud platform. To complement Charmed OpenShell, Canonical has engineered an open source LXD driver that connects the OpenShell gateway directly to MicroCloud's virtualization layer. By using LXD's system containers, the driver provisions isolated sandboxes at the container boundary for every agent execution. This architecture allows OpenShell to dynamically interact with MicroCloud storage, networking, and physical GPU passthrough primitives, while helping to guarantee that unverified binaries or compromised agent code remain safely contained. "Autonomous agents represent a fundamental shift in software execution, moving from predictable scripts to self-directed systems," said Cindy Goldberg, Vice President of Cloud and Silicon Partnerships at Canonical. "Deploying these agents at scale demands infrastructure that provides strict governance without stalling developer velocity. Through our collaboration with NVIDIA, Canonical is delivering a complete open source stack - from developer snaps to Charmed OpenShell on MicroCloud - giving organizations the safety, scale, and operational control needed to run AI agent fleets with confidence." Availability and resources. To install OpenShell locally on your Ubuntu machine run sudo snap install openshell To learn more about Charmed OpenShell consult the documentation, or contribute via the GitHub repository. About Canonical. Canonical, the publisher of Ubuntu, provides open source security, support, and services. Its portfolio covers critical systems, from the smallest devices to the largest clouds, from the kernel to containers, from databases to AI. With customers that include top tech brands, emerging startups, governments, and home users, Canonical delivers trusted open source for everyone.

Mir 2.30 released now with Rust required, mir roadmap published. Canonical today released Mir 2.30 as the newest release of this Wayland compositor library for building out different shells and other functionality from smart mirrors to other consumer devices. Mir 2.30 now explicitly requires a Rust compiler for building Mir where as previously it was considered optional. Mir is now using the Rust evdev input platform code. The other motivation is that Mir is now strictly targeting Ubuntu 26.04 LTS where there is Rust mandated. So thanks to upping their OS target focus of Ubuntu 26.04 and newer, Rust can be safely and always assumed for Mir purposes. The "MIR_ENABLE_RUST" build option is now removed and Rust must always be present for building Mir. Mir 2.30 also adds support for the Wayland wl_fixes protocol, fixes various bugs, and also brings various documentation improvements. Mir 2.30 can be downloaded from GitHub. With this release added to the Mir 2.30 documentation is a public roadmap. Among the roadmap plans include using a hardware plane for video playback, support for building a Mir-based compositor entirely in Rust, stabilizing the mir-shell Wayland extension, designing a long-term Mir rendering pipeline, Mir shell systemd integration, and security improvements.