Skip to main content

Roadmap

We're not stacking features. We're building the layers of a foundation, in order. Each milestone is usable on its own, and each release is testable — host-first, then proven in QEMU. The early layers below aren't promises; they already boot.


Where we are now

It all runs in QEMU — emulated RISC-V, not yet a board on your desk. That's the honest scope. But it's real code you can clone and boot, not a slide deck.

Kernel & core services. The NEURON microkernel boots with capability-based IPC, process isolation, W^X memory, and a 14-syscall surface — deny-by-default, enforced. On top run the core services: service manager, policy engine, bundle manager, virtual filesystem, keystore, logging. See the architecture.

A UI on the GPU. A full input-to-output loop driven by real interrupts — a kernel IRQ timer, and a GPU command ring buffer with IRQ-driven completion. Input comes back as motion with a fast path for the cursor; rendering runs on the GPU over virtio-gpu, accelerated with virgl, with opacity, blur, and shadow. Windows are real applications. See We Made the Screen Light Up.


What's next

The screen is on. Next comes a real way into the system, then the parts that make it a joy — first to build on, then to live in.

Login & session. A real entry point — authenticated sessions built on the identity and keystore services.

Shell infrastructure & design primitives. The foundation a desktop is assembled from: the core surfaces, components, and layout contracts the shell and apps will share.

The Nexus DSL. Our own language for describing interfaces and how they bind to the services beneath them — declarative, testable, and the piece that makes Open Nexus fun to build on.

A new shell, written in the DSL. Not hand-wired this time — assembled from the primitives and expressed in our own language. A shell and launcher that feel like home. The piece that makes Open Nexus fun to use.


Further out

Secure device platform & first deployments. Boot lifecycle, policy enforcement, and update paths hardened for industrial HMIs, kiosks, and public-sector endpoints — where security and auditability differentiate.

Broader device classes. Expand from focused deployments toward laptops, mobile, and wider ecosystems as the architecture and partner base mature. One OS. Many Devices. In the right order.

Enterprise & advanced. Device management, compliance, distributed systems (DSoftBus), ML acceleration — for organizations that will trust this platform at scale.


How we ship

Slice by slice. From a booting kernel → to secure specialized deployments → to broader device classes. Not rewrites for sport. Not roadmaps on slides. Just code, progress, and proof you can run yourself.