Skip to content

Repository files navigation

LaplacianOS

LaplacianOS is a Rust operating system with a capability-oriented kernel, native typed temporal graph, predictive operating-system control, and a built-in SCCE cognitive runtime. It boots through a custom UEFI loader and runs native ring-3 services and applications on a GPU-composited desktop.

Architecture

Layer Responsibility
UEFI loader Loads the kernel and transfers typed boot information.
Kernel Scheduling, memory, IPC, capabilities, drivers, bounded graph mutation, and telemetry.
Graph runtime Typed nodes and edges, temporal attenuation, deterministic PowerWalk, provenance, causal evidence, spectral state, and predictive telemetry.
Ring-3 services graphd, modeld, trainerd, artifactsd, sysd, servicemgr, compositor, networking, and supporting services.
SCCE runtime Document ingestion, retrieval, entity correlation, graph reasoning, conversation state, tool planning, and auditable outputs.
Applications AI Console, Predictive System Monitor surfaces, Terminal, Files, Settings, Editor, Browser Lite, and 3D applications.
SDKs and tools Application, UI, GL, graph, tool, WASM, packaging, installation, and verification tooling.

LaplacianOS has two distinct intelligence domains that share the graph, provenance, event, and capability substrate:

  • Operating-system intelligence processes live telemetry to forecast system state and propose bounded, reversible actions.
  • SCCE cognition processes documents, entities, relations, conversations, and tool results for local retrieval, reasoning, synthesis, and tool use.

Heavy corpus processing, long-running prediction, and model orchestration run in ring 3. Kernel paths remain bounded and must not block on prediction.

Repository layout

  • kernel/: kernel, graph substrate, cognitive bootstrap, and hardware/runtime integration.
  • boot/: UEFI boot loader.
  • userspace/: native services, compositor, shell, and protected ring-3 runtime.
  • apps/: native applications.
  • sdk/: application, UI, GL, graph, tool, WASM, and hashing SDKs.
  • tools/: package, image, benchmark, cryptographic, and verification tools.
  • scripts/: build, QEMU, release, and verification entry points.
  • docs/: architecture, ABI, operational, and validation documentation.

Build requirements

  • Rust nightly as pinned by rust-toolchain.toml.
  • rust-src, llvm-tools-preview, rustfmt, and clippy.
  • QEMU and OVMF for x86_64 system testing.
  • PowerShell 7+ for the Windows build and verification scripts.

Quick checks

cargo check -p laplacianos-kernel
cargo check -p laplacianos-trainerd
cargo check -p laplacianos-artifactsd
cargo check -p laplacianos-sysd
cargo check -p laplacianos-ai-console
cargo check -p laplacianos-appstore

For the repository verification gate:

powershell -ExecutionPolicy Bypass -File .\scripts\verify.ps1 -SkipBoot

To boot the x86_64 system in QEMU:

.\scripts\boot-windows-qemu.ps1

Mathematics

The temporal graph and predictive runtime use explicit, testable operators with documented units and assumptions. The name Walsh does not imply a Walsh-Hadamard transform. LaplacianOS does not use the removed type-space FWHT interpretation. Authoritative mathematical implementations must be traceable to their source equations and independent numerical validation.

License

LaplacianOS is source-available under the PolyForm Noncommercial License 1.0.0.

The public license permits qualifying noncommercial use, study, modification, and redistribution under its terms. It does not authorize commercial use.

Any commercial use requires a separate written license obtained directly from Sarah Walsh. Commercial licensing inquiries may be made through LinkedIn or the LaplacianOS GitHub issue tracker.

Contacting the licensor does not itself grant permission. Commercial rights are granted only through a separate executed written agreement.

See:

  • LICENSE
  • NOTICE
  • COMMERCIAL-LICENSING.md

About

A ground-up Rust OS with UEFI bootloader, capability-secured microkernel, GPU-composited desktop, and an on-device cognitive engine. No LLM. Deterministic inference.

Resources

Contributing

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages