Projects

What We're Building

From an AI-native language and microkernel to a systems language for the CLR, here's what's in progress across Deepcomet's branches.

Core Projects

The languages, kernels, and systems that make up the Deepcomet stack.

Idea Stage

Aurelia

Deepcomet AI

An AI-native language with a Rust/Python-hybrid syntax, built on a C++23, MLIR, and Go toolchain, with a single stable C ABI / gRPC boundary between components.

.aulC++23MLIRGo
Active

Zenith Kernel

Deepcomet AI

An ARM64/NPU microkernel, currently incubating under the name Perception Kernel while it works toward a first experimental release.

ARM64NPUMicrokernel
On Hold

SkyOS

Deepcomet AI

A civilization-scale operating system, on hold until Zenith Kernel stabilizes and reaches its first experimental release.

OSCivilization-Scale
Active

Voyage

Deepcomet Tech

A Swift-inspired systems and application language built directly on the CLR — reentrant actors, compile-time safe concurrency, and native C# interop. Full spec complete; parser at 248 passing checks.

CLRSwift-InspiredActors
Idea Stage

S#

Deepcomet AI

A scientific computing language for the .NET runtime, inspired by modern Fortran, built with NativeAOT — the original idea that gave rise to the rest of the Deepcomet AI stack.

.ssharp.NETFortran-Inspired
Idea Stage

Ferrum OS

Deepcomet AI

A lightweight modern OS concept based on Android Froyo, with the Java application layer replaced by a single C# runtime — no mixed-runtime FFI boundary.

Android-BasedC#Lightweight

Scientific Packages

Smaller tools and libraries supporting Deepcomet AI's scientific computing vision.

OpenUnits

An orbital-mechanics CLI built for Node.js, part of Deepcomet AI's scientific tooling.

ThermoUnits

A zero-runtime-dependency Python library for thermodynamics, with IAPWS-IF97 water models and Shomate air properties, backed by a 43-test suite.

Astronomy Visualizer

A planned cross-platform 3D astronomy visualizer for Deepcomet Science, exploring Godot 4 or an Electron + Python/Astropy/SpiceyPy stack backed by JPL HORIZONS data.

Interested in Collaborating?

We're always looking for engineers and researchers to help build these projects out.

Reach Out