Aurora Nano Studio
Architecture, solvers, and validation for Aurora Nano Studio: a physics-grade atomistic workbench with cloud, desktop, and offline compute. This paper covers the modifier pipeline, the swappable compute layer (inline, desktop, HPC), file-format coverage, and the validation harnesses shipped with the platform.
Architecture & compute layer
Aurora Nano isolates studios from compute providers behind a typed resolver: inline (browser JS), local worker (desktop / offline), and cloud. Studios never import backends; a single call site (nanoCompute.*) resolves the best available provider for each operation.
Full detail in the embedded whitepaper.
Modifier pipeline
20+ chained, cached, deterministic modifiers — coordination, RDF, Voronoi, CNA, dislocations, Wigner-Seitz, interface relaxation, powder diffraction, Madelung, hydrogen bonds, bond-valence sums, magnetic mCIF. Each modifier is a pure function of atoms + parameters and memoized in an LRU.
Full detail in the embedded whitepaper.
File format coverage
Native parsers for XYZ, mmCIF, PDB, GRO, LAMMPS, DUMP, POSCAR, mCIF and writers for glTF, VRML, M3D, JSON bond graphs, and Python script export.
Full detail in the embedded whitepaper.
Validation harnesses
Vitest suites cover crystallography (lattice detection, Bravais classification), bond inference (Cordero / Pauling), and modifier correctness; visual golden tests use Playwright + pixelmatch on fixed scenes.
Full detail in the embedded whitepaper.
Trouble viewing? Open the white paper in a new tab, or download the PDF.
Cite this document
When referencing Aurora Nano Studio in academic or industry work, please cite the technical white paper below.
@techreport{aurora_nano_studio_whitepaper,
title = {Aurora Nano Studio — Technical White Paper},
author = {{Aurora Simulations}},
year = {2026},
url = {https://simaurora.com/whitepaper},
note = {Physics-grade atomistic workbench with swappable compute layer}
}