The Quantum Metal Ecosystem

Quantum Metal is the open-source chip-design layer for superconducting quantum hardware. A growing community of tools builds on top of it, extends it with new simulation backends, and plugs into it for quantization, visualization, and parameter discovery.

This page maps that ecosystem so you can find the right tool for what you’re trying to do — and so we can highlight the projects that already choose Quantum Metal as their substrate.

If you maintain a project we should list here, please open an issue or ping us on Discord.

Built on Quantum Metal

These projects use Quantum Metal directly — they consume QDesign objects, extend the QComponent library, or treat Metal as their foundation:

Project

What it does

License

SQuADDS (LFL-Lab @ USC)

Validated qubit-design database + physics-based parameter interpolation. Search the DB, interpolate to your target Hamiltonian, generate the design in Quantum Metal. Published in Quantum journal (Sept 2024). Includes an MCP server for AI agents.

MIT

SQDMetal (SQDLab @ UQ)

Simulation wrapper for QDesign → AWS Palace / COMSOL workflows. Covers eigenmode, capacitance, driven, inductance simulations end-to-end. Accepts Quantum Metal QDesign objects directly.

Apache 2.0

ML_qubit_design (Fermilab + Northwestern)

ML-based inverse design — predicts Quantum Metal design parameters from target qubit, resonator, coupler, and Hamiltonian properties using multi-layer perceptrons. Notebook-driven research project.

see repo

pypalace (Northwestern)

Python toolkit for AWS Palace with Quantum Metal gmsh export, JSON config builders, LOM analysis utilities.

see repo

This list grows with the community. If you’re building on Quantum Metal in a research project, an internal tool, or a startup, we’d genuinely love to know — open an issue and we’ll add you.

Solvers Quantum Metal integrates with

These are the simulation backends Quantum Metal drives via its renderer protocol. Each is its own project; we wrap, we don’t fork.

Solver

Role

Integrated via

AWS Palace (AWS Center for Quantum Computing)

Maxwell solver — eigenmode + driven + electrostatic + magnetostatic + AMR. Apache 2.0.

Roadmap (see ROADMAP.md)

Ansys HFSS / Q3D

Industrial-grade EM solvers. Closed-source, requires AEDT license.

[ansys] extra → pyaedt

Elmer FEM

Open-source FEM solver. Today: capacitance for LOM analysis.

[mesh] extra (gmsh + Elmer external binary)

gmsh

Workhorse mesh generator (used by Elmer today, Palace tomorrow).

[mesh] extra

Quantization & analysis libraries

After simulation produces fields / S-parameters / capacitance, these libraries turn the results into qubit physics:

Library

Role

Integration

pyEPR-quantum

Energy-Participation-Ratio quantization from HFSS field data. The math we use to turn eigenmodes into Hamiltonians.

[ansys] extra

scqubits

Closed-form qubit-spectra and circuit-Hamiltonian library.

Base deps

QuTiP

Quantum dynamics — time evolution, master equations, etc.

Base deps

CircuitQ

Quantization of arbitrary superconducting circuits.

Aware, not yet integrated

Visualization

For viewing simulation outputs (Palace / Elmer field data, mesh files):

  • PyVista — Python ParaView wrapper, MIT.

  • ParaView — the underlying renderer, BSD.

The [mesh] extra and Palace renderers emit .vtu / .pvtu files these tools open natively.

How the layers fit together

A typical workflow walks through several of these projects:

  1. Discover a candidate design via SQuADDS (search the DB, interpolate to your target parameters) — or hand-design from scratch using Quantum Metal’s QComponent library.

  2. Build the QDesign in Quantum Metal — instantiate QComponents, place + route, set options.

  3. Simulate via the renderer of your choice — Ansys via [ansys], open-FEM via [mesh] + Elmer, or AWS Palace via SQDMetal / pypalace (coordination underway — see ROADMAP.md).

  4. Analyse — EPR (pyEPR), LOM (LOManalysis), spectra (scqubits), dynamics (QuTiP).

  5. Export to GDS via the built-in QGDSRenderer; view in KLayout or hand off to fab.

Some users also close the loop with ML inverse design — ML_qubit_design trains on Quantum Metal simulation data to predict QDesign parameters from target qubit properties, enabling fast parameter exploration without running full EM simulations.

Why we ecosystem-map instead of building it all

Quantum Metal’s job is to be the best chip-design layer it can be — QComponent library, renderer protocol, headless qm.view(), GDS export, the lite-by-default install. We’re explicitly not the right place to reimplement a Maxwell solver, a qubit-design database, an ML inverse-design framework, or a photonic-FEM library. Other projects do those well.

The community wins when these projects coordinate at the edges — shared QDesign formats, compatible meshing conventions, mutual docs links — rather than competing for the same maintainer-hours.

If you want to help shape that coordination, see the Adoption / DevRel section in ROADMAP.md or reach out on Discord.