A graph-native operating-system kernel
for x86_64 bare metal.

OverviewProduct

GOS walks away from the traditional loader-first / procedure-first OS architecture and puts the graph at the heart of the kernel — an experimental operating system for x86_64 bare metal.

Nodes, edges, vectors, capabilities, mount, and use are public execution concepts. Everything is written in Rust, threading vertically through four layers: Hypervisor → Runtime → Supervisor → Plugins.

Four-layer ArchitectureArchitecture

① Hypervisor — Bootstrap Layer

Performs only minimal bootstrap and CPU initialization. A "bring it up and hand it off" philosophy that maximises the freedom of the layers above.

② GOS-Runtime — Graph Execution Layer

The core runtime managing plugins, nodes, edges, activation, and routing. The structure of the graph itself is the flow of execution.

③ GOS-Supervisor — Management Layer

Controls modules, domains, and the system service cycle. Services are treated not as "processes" but as activated regions on the graph.

④ Native Graph Plugins

Shell, Cypher, AI, CUDA, networking, and I/O — all implemented as graph plugins. The Cypher v1 subset lets you query kernel state directly.

CapabilitiesFeatures

Graph shell

Cypher v1 subset — query kernel state directly with MATCH / RETURN

CUDA bridge

GPU compute abstracted as graph nodes

Networking

Declarative network state through status nodes

Mount system

mount resources (e.g. clipboard) as graph elements

Capabilities

Permissions / resources as first-class concepts via use semantics

Tech StackEngineering

Language

Rust (nightly + rust-src + llvm-tools-preview)

Target

x86_64 bare metal

Dev environment

QEMU (virtual machine), PowerShell 7 (build automation)

Docs

Chinese & English — architecture, CLI commands, bare-metal install

Status

Substrate stabilization — higher-level features expand after native module execution lands

Why a Graph?Philosophy

Traditional operating systems are built on the abstraction of "process × memory." But modern systems are built on "service × topology."

By treating the graph as a first-class citizen from the very start, GOS aims to handle microservices, GPUs, networking, and AI under one set of semantics. Instead of "spawn a process," activate a node; instead of "connect a pipe," draw an edge.

Open SourceCommunity

GOS Kernel is developed in the open on GitHub.
Explore the architecture, CLI commands, and bare-metal setup docs.

Get in Touch

For technical discussion or collaboration, we'd love to hear from you.