Quantum + Agentic AI

SuperQuantX

The agentic layer for Quantum AI

An open software layer across quantum SDKs, with a classical baseline for every result. Open source Python SDK for PennyLane, Qiskit, Cirq, Amazon Braket and pytket.

superquantx.dev GitHub 📚 Documentation PyPI

Apache 2.0. Version 0.3.0, alpha. Experimental research software.

SuperQuantX 0.3.0, local simulator

$uv add "superquantx==0.3.0"

✅ + superquantx==0.3.0

$python

>>> import superquantx as sqx

>>> backend = sqx.SimulatorBackend(max_qubits=2)

>>> circuit = backend.create_circuit(2)

>>> circuit.h(0).cnot(0, 1) # Bell state

>>> result = backend.execute_circuit(circuit, shots=1024)

>>> print(result["counts"])

{'00': 518, '11': 506}

⚛️ Only |00⟩ and |11⟩ appear: the two qubits are entangled

Counts vary run to run. Local simulator, no account needed.

Quantum

⚛️ What is SuperQuantX?

SuperQuantX is the agentic layer for Quantum AI: an open software layer across quantum SDKs, with a classical baseline for every result. It is an open source, experimental Python SDK for quantum algorithms and quantum-inspired agents, so you can see whether the quantum part helped.

🔬 The Vision

One consistent interface across quantum frameworks, with a classical comparison built in.

Work across PennyLane, Qiskit, Cirq, Amazon Braket, pytket and a built-in simulator
Switch backends by changing one argument
Accelerate quantum machine learning research
Make quantum computing accessible to AI researchers
Report a classical baseline beside every quantum result

🚀 Why Now?

The quantum computing landscape is fragmented across different frameworks, each with unique strengths but incompatible APIs.

SuperQuantX solves this with:

One consistent API across supported frameworks
Research-first design for experimentation
VQE, QAOA, QuantumSVM and agents, each with a classical reference
An optional MCP server with six read-only tools for AI assistants

No quantum advantage claimed. The baseline tells you what helped.

🎬 See SuperQuantX in Action

A walkthrough of SuperQuantX running on a local simulator

Launch Demo

•

Local Simulator

•

Real-time Examples

✨ Core Features

What ships in the open source SDK today

🔗 Unified API

One interface for several quantum backends. Switch frameworks by changing the backend argument

🎯 Quantum Agents

Portfolio and classifier agents that compare their answer to the classical optimum

🚀 Multi-Backend Support

PennyLane, Qiskit, Cirq, Amazon Braket, pytket and a built-in simulator, through optional extras

📊 Advanced Algorithms

VQE, QAOA and QuantumSVM, plus compare_to_classical for cross-validated benchmarks

🛠️ Developer Friendly

Comprehensive documentation, examples, and educational resources

⚖️ Classical Baselines

Every result arrives next to a classical reference, so claims stay checkable

🧪 Research-First Design

Reproducibility, flexibility, and experimentation built into every feature

🤖 MCP Server

Optional MCP server with six read-only tools for AI assistants, simulators only by default

💻 Command Line

A superquantx CLI to list backends and run example algorithms from the terminal

🎯 Supported Quantum Backends

Install the extra for the framework you use. Hardware access depends on that provider and your own account

🍃

PennyLane

Xanadu

Differentiable circuits
ML library integration
Optional extra

🔷

Qiskit

IBM

Qiskit circuits
Transpilation via Qiskit
Optional extra

🔴

Cirq

Google

Cirq circuits
Cirq simulators
Optional extra

🟡

Amazon Braket

AWS

Braket SDK circuits
Local Braket simulator
Optional extra

⚪

pytket

Quantinuum

pytket circuits
Circuit optimization passes
Optional extra

🌊

D-Wave Ocean

D-Wave

QUBO and Ising problems only
Hardware needs a Leap token, untested
Optional extra

💻

Simulator

Built in

No extra dependencies
Runs every example
Default for MCP tools

🚀 Get Started in Minutes

Each example runs on the released 0.3.0 SDK with the built-in simulator

Your First Quantum Circuit

# Bell state on the built-in simulator

from superquantx import SimulatorBackend

backend = SimulatorBackend(max_qubits=2)

circuit = backend.create_circuit(2)

circuit.h(0).cnot(0, 1)

result = backend.execute_circuit(circuit, shots=1024)

print(result["counts"])

# {'00': 518, '11': 506} (varies per run)

Benchmark Against Classical

# QuantumSVM against classical models

import superquantx as sqx

from sklearn.datasets import make_moons

X, y = make_moons(n_samples=100, noise=0.2, random_state=0)

report = sqx.compare_to_classical(

X, y, sqx.QuantumSVM(backend="simulator"), cv=5

)

print(report)

# Cross-validated accuracy for the QSVM and classical

# baselines. On this toy data the best classical model

# matches it (no significant difference, p=0.54).

QAOA MaxCut

# QAOA MaxCut with the classical optimum

import numpy as np

import superquantx as sqx

W = np.array([[0, 1, 1, 0], [1, 0, 1, 1],

[1, 1, 0, 1], [0, 1, 1, 0]])

result = sqx.QAOA(backend="simulator", p=2, seed=7).solve(W)

print(result["cut_value"], result["classical_optimum"])

print(result["approximation_ratio"])

VQE for H2

# VQE for H2, checked against exact diagonalization

import superquantx as sqx

vqe = sqx.create_vqe_for_molecule("H2", backend="simulator")

print(vqe.find_ground_state())

print(vqe.exact_ground_energy())

# Both about -1.857 on the simulator

Try the Full Quickstart Guide

🔬 Research Applications

Areas the SDK supports for research and learning

Quantum Machine Learning

QuantumSVM with quantum kernels, benchmarked against classical models

Quantum Optimization

QAOA and VQE, each compared with a classical optimum or exact answer

Quantum Agents

Portfolio and classifier agents that report the classical answer alongside their own

Hybrid Algorithms

Variational loops with a classical optimizer driving the quantum circuit

AI Assistants via MCP

Six read-only MCP tools so assistants can run small simulated circuits and demos

Educational Research

Small, readable examples for students learning quantum algorithms

Try SuperQuantX

Install the SDK, run the examples on the simulator, and compare with the classical baseline

Open Source

Apache 2.0 license, developed in the open on GitHub

Well Documented

Guides and API reference on the docs site

Alpha Release

Version 0.3.0. APIs may change between releases

Get Started on GitHub Visit superquantx.dev 📚 Read the Documentation

Quick Start: Install with uv add "superquantx==0.3.0" (or uv pip install "superquantx==0.3.0" in an existing environment) from PyPI. Python 3.11 or newer.