Omarchy Quattro: when a distro goes “everything inside” on AI and Hyprland

Omarchy launches Quattro, a edition that makes AI coding agents first-class citizens inside a distro built on Hyprland, a tiling window manager that demands dozens of keyboard shortcuts. The result is a powerful workspace for developers willing to pay the learning-curve price, but the entry barrier is high for anyone expecting a conventional desktop.

AI as built-in, not an add-on. According to the official documentation, Omarchy does not pick a preferred agent but “pre-wired” every central CLI of coding agents as a “lazy-loaded launcher” managed through mise, meaning nothing downloads from the network before the first run. In a practical test, configuring Gemini opened a browser for verification that got trapped behind the overlay of the installer wizard; only launching a third-party app that forced a window re-tiling let the process finish. Google verification failed every time and required a manual API token. After that the agent acted as a command-line tool that enables chat with the chosen service, including sandbox support.

Hyprland: a keyboard puzzle. Hyprland itself is not beginner-friendly; it requires memorizing shortcuts and manual window management, and Omarchy does not try to hide that. The tester, familiar with Hyprland, reported moments of “hmm” when the automatic tiling behaved unexpectedly even for him. This is not a distro for the average user; the declared target audience is developers, and the choice of Hyprland reflects that without apology.

Local and cloud agents side by side. In addition to cloud agents, the system offers installation of LM Studio via Applications > Install > AI. LM Studio provides a GUI for downloading models and running them locally, an advantage over Ollama which installs only a CLI. The system lets LM Studio run alongside Omarchy’s default agent, so a local model can be used in a graphical interface while a cloud agent runs in the CLI simultaneously.

Automatic crash monitoring and subscription tracking. A top-bar icon appears when the system detects use of coding agents; the icon aggregates subscription data, plan, utilization percentage, weekly limits and daily token usage per model. At the system level, systemd-coredump is monitored by AI: when an application or service segfaults, a “Process crashed” alert shows. Clicking it sends the core dump to the default agent together with a built-in skill named diagnose-crash, which guides the agent to collect dump information and decide whether to report. Additional agent skills enable actions such as tail, extending the AI’s ability to operate on logs and the environment in real time.