If you’ve ever sat at a computer with three browser windows open, a chat app blinking, and a dozen files scattered across folders, trying to coax an AI assistant into helping you with something that should be simple, you know the feeling: the tools feel brilliant and broken at the same time. You paste in context, explain what you need from the beginning, and hope the AI remembers what you just showed it. But it doesn’t. It’s like having a brilliant coworker with amnesia. A new operating system called Kora wants to change that by putting AI not on the sidelines, but at the very heart of the machine. Kora is built to tap into your data and context at the operating-system level, meaning it can see the files, messages, and memory living on your own device. “All your context, all your memory is on your machine,” says Franck De Girolami, one of Kora’s cofounders. That simple idea changes the relationship between a person and their computer. Imagine designing a project with an OpenAI model inside Kora’s interface, then switching to Anthropic’s tools to deploy code, and later leaning on a local model to save money on AI tokens. The AI stops being a website you visit and starts becoming a part of the operating system itself, like the cursor or the window manager. It still feels like a computer, but it’s one that actually remembers you.
Kora is not a skin on top of Windows or macOS. It’s a standalone operating system, built on the Linux kernel, that runs directly on your hardware. That might sound like a heavy lift, but the company has already partnered with Qualcomm to bring Kora to Snapdragon PCs next year. It also supports x86-64 architecture, so if your current computer runs on an Intel or AMD chip, there’s a good chance you’ll be able to install Kora and give it a try. But the makers of Kora aren’t naive enough to think you’ll want to give up every app you rely on overnight. To bridge the software gap, the operating system includes built-in compatibility layers that let users run standard Windows applications. The goal is to let you keep your old familiar tools while gradually trying the new ones. The company is also building its own native apps for the basics every desktop user expects—contacts, calendar, email—alongside a software development kit, or SDK, that invites third-party developers to build for the platform. The visual design leans on language we already understand: along the bottom of the home screen is a dock, exactly where you’d expect it on a desktop OS. That’s where you switch between email, web browsing, and launching games. In other words, Kora is trying to be both familiar and radically different at the same time. You get the comfort of the old world and the promise of a new one, without having to relearn how to use a computer from scratch.
The heart of Kora is something the startup calls a “neural file system,” and it’s here that the operating system really starts to feel less like software and more like a thinking companion. On traditional computers, your files are stored by name and location. You have to remember whether you saved that document in Documents, Desktop, or some mysterious folder called “Final_Final_v3.” Kora flips that model. It indexes file data by meaning, not just by name, so you can ask a question about the broad context of your files by typing into an OS-level search bar, without needing to recall a specific file name. You could search for a business contract saved somewhere on your device based on a small detail you remember—maybe the name of the client’s assistant, a phrase from the opening paragraph, or the date you met. Kora uses its understanding of context to find it. The user experience feels a little like messaging a chatbot or an AI agent, because the search isn’t just a search; it’s a conversation. You’re not querying a database; you’re asking the computer itself to remember something. For anyone who has ever lost hours digging for a file, this idea is deeply appealing. It’s the difference between organizing your life by folder names and organizing your life by what things actually mean. And because all of this memory is local, on your machine, there’s a privacy angle too: the system isn’t required to send your most personal documents to the cloud to understand them. The computer becomes less of a filing cabinet and more of an extension of your own memory.
Kora also has a surprising origin: it grew out of Playtron, a startup founded by McMaster that launched its own gaming operating system, GameOS, last year. GameOS was built primarily for portable handhelds like the Steam Deck, with PC support as well. That gaming DNA now lives inside Kora as a component called GameFrame, which lets users run PC games as long as their hardware is powerful enough. This is not just a quirky side feature; it’s central to Kora’s identity. The team wants to attract not only productivity junkies but also PC gamers and power users—people who expect their computer to be fast, flexible, and capable of handling demanding creative work. Setting up a gaming session should be as easy as opening your email, and Kora is trying to make that happen without making you feel like you’ve traded one operating system’s problems for another’s. “That’s always our thing—next-generation creators who are using these tools daily to build new applications and services, but they want everything else that they can get from a Windows PC or what’s currently available on macOS,” McMaster says. That quote captures exactly who Kora is for. It’s not aimed at people who just check a few websites; it’s aimed at people who build, create, play, and live on their machines. They want a modern AI-first environment, but they don’t want to give up the raw capability they already rely on. By folding the gaming engine into the same system as the neural file system, Kora bridges two worlds that often feel disconnected: the world of play and the world of work.
Building a new operating system is a massive ambition, and Kora’s team knows that the biggest hurdle isn’t just making something beautiful—it’s making something that fits your life. That’s why compatibility is so important. The built-in layers that allow standard Windows apps to run mean that you don’t have to abandon your favorite software just to try something new. The native apps for contacts, calendar, and email give you a solid foundation, but the SDK is the real invitation: it’s a signal to developers that Kora isn’t just a closed experiment, but a platform where they can build and share. The partnership with Qualcomm and the promise of Snapdragon PCs adds another layer of momentum, suggesting that Kora will be more than a niche hacking project. But perhaps the most compelling piece is the idea of a system that makes AI useful in ordinary moments. When it’s easy to switch between an OpenAI model, an Anthropic model, and a local model—depending on what you’re doing and how much you want to spend—AI becomes less of a single magic box and more of a flexible utility. You choose the right tool for the job, the same way you might choose a favorite text editor or web browser. And because Kora has access to your context, it can make intelligent suggestions without you having to explain your whole life story. If you’re working on a project, it knows the relevant files are nearby; if you need to find a message from a colleague, it knows where to look. The speed bumps that plague today’s AI assistants—the endless copying and pasting, the lost context, the need to rephrase every request—are exactly the speed bumps Kora is designed to remove.
In the end, Kora’s vision is about something even deeper than convenience. It’s about making computers that feel less like strangers to us. Modern desktop operating systems were designed long before artificial intelligence existed, and they still ask us to think the way machines think: remember file names, remember folder paths, remember the exact name of that application you installed two years ago. Kora asks what would happen if the machine learned to think a little more like us. If you could ask questions instead of performing searches. If the system could remember not just what you saved, but why you saved it and how it fits into what you were doing. The promise of the neural file system is that your computer can become a true partner, holding on to the ephemeral details that slip out of your mind while you’re busy doing actual work. And by pairing that with gaming, with Windows compatibility, with native apps, and with an SDK, Kora is trying to make that future practical, not just theoretical. It is a huge promise, and there will certainly be challenges ahead: adoption takes time, developers need reasons to commit, and turning AI from a novelty into an operating system requires trust. But the direction feels right. For anyone who has ever felt that their computer was too intelligent in the abstract and too dumb in the details, Kora offers a picture of a different kind of desktop—one where your documents, your memories, and your tools all live in the same place, and where the machine finally seems to understand what you mean.
