DOCUMENTATION

Software that makes quantum computing cheaper, safer, and predictable.

Quantum hardware is expensive and error-prone. QubitBoost detects hardware failures early and prevents wasted compute — so quantum workloads only proceed when the hardware is actually working.

The platform provides three layers of execution governance, applied before, during, and after quantum jobs run.

THE QUBITBOOST EXECUTION STACK

TOMOGATE

Pre-Flight ValidationHARDWARE VALIDATED

TomoGate certifies that quantum states meet required fidelity thresholds before expensive downstream workloads begin.

  • •Orders-of-magnitude fewer circuits and shots than full tomography
  • •Provably equivalent go/no-go decisions
  • •Fails closed when confidence is insufficient

LIVEGATE

Early-Flight Budget ProtectionHARDWARE VALIDATED

LiveGate monitors execution in real time and aborts runs early when statistical evidence shows they are likely to fail. It applies to any runnable quantum workload (QEC, VQE, QAOA, calibration, benchmarking), without modifying circuits, hardware, or user code.

  • •Confidence-interval-based abort decisions
  • •Never increases total cost
  • •Worst-case behavior: no abort (job runs normally)

SAFETYGATE

QEC Trust CertificationSTIM VALIDATED

SafetyGate is a trust layer for quantum error correction that determines whether decoder corrections should be applied or suppressed.

  • •Up to 62% fewer false alarms compared to apply-all strategies
  • •Supports code distances d=3 through d=9
  • •Fail-closed design: suppresses correction when confidence is low

GUARDGATE

QAOA Solution Quality GateHARDWARE VALIDATED

GuardGate validates QAOA solutions before deployment, catching invalid or low-quality results that would otherwise waste downstream compute. Enforces approximation ratio thresholds and constraint satisfaction.

  • •Catches constraint violations before costly deployment
  • •Configurable quality thresholds per problem type
  • •Three-tier decisions: ALLOW, WARN, ABORT

OPTGATE

QAOA Optimization ControlHARDWARE VALIDATED

OptGate adaptively controls QAOA optimization, dramatically reducing QAOA shot budgets while maintaining solution quality — decisions that today are made by expensive human trial-and-error.

IBM FEZ HARDWARE VALIDATED (FEB 2026)

117x+

FEWER SHOTS

128

vs 26,000+

100%

WIN RATE

Tested on complete-graph MaxCut problems from 6 to 20 qubits (80 trials). OptGate uses 128-448 shots vs 26,000+ for classical optimizers (COBYLA, CMA-ES, Nelder-Mead) — a 117-208x reduction on IBM. Trained on real calibration data from 4 vendors (IBM, Rigetti, IonQ, IQM) across 8 backends. Hardware validated on IBM Fez, Torino, and Marrakesh.

  • •117-208x shot reduction vs COBYLA, CMA-ES, Nelder-Mead (hardware validated)
  • •4 vendors, 8 backends: IBM, Rigetti, IonQ, IQM with vendor-specific models
  • •Trained on 146 real quantum hardware calibration snapshots
  • •Compatible with any error suppression layer (Fire Opal, QESEM)

CHEMGATE

VQE Cost OptimizerHARDWARE VALIDATED

ChemGate accelerates variational quantum eigensolvers by reducing the number of evaluations needed, while maintaining chemical accuracy. A built-in safety mechanism prevents accuracy loss. Hardware validated on IBM Fez.

  • •32-42% fewer evaluations with zero accuracy loss (IBM + Rigetti validated)
  • •Fail-closed: automatically recovers when optimization is too aggressive
  • •Validated on drug-sized molecules (caffeine 8q, warfarin 16q)

SHOTVALIDATOR

QEC Shot Quality FilterHARDWARE VALIDATED

ShotValidator uses redundant syndrome copies to identify and discard corrupted QEC shots in real time. Feeds into ExecutionMonitor for automated abort decisions. Hardware validated on IBM Fez with 216K shots across 27 configurations.

  • •100% error capture in 6/9 error-producing QEC configurations
  • •74% mean error reduction across all configurations
  • •Integrates with SafetyGate and ExecutionMonitor pipelines

DESIGN PHILOSOPHY

QubitBoost does not promise quantum advantage. It does not modify algorithms. It does not compete with SDKs, notebooks, or cloud providers.

Instead, it provides execution governance:

1
when to proceed
2
when to abort
3
when to defer
4
when results can be trusted

All decisions are statistically justified, auditable, and reversible.

VALIDATION & SCOPE

QubitBoost has been validated across multiple commercial backends using real hardware execution, controlled simulations, and statistically auditable decision logic.

The system is designed to integrate with existing quantum platforms, not replace them.

GETTING STARTED

Install the SDK

terminal

$ pip install qubitboost-sdk

PYTHON
from qubitboost_sdk import LiveSession

with LiveSession(backend="ibm_fez", doom_threshold=0.7) as session:
    result = session.run(circuit, shots=10000)
    if result.aborted:
        print(f"Saved {result.budget_saved_percent:.1f}% budget")
    else:
        print(f"Doom rate: {result.doom_rate:.3f}")

API Reference

Python quickstart, Qiskit integration, endpoints, and rate limits.

Jupyter Notebooks

9 interactive tutorials — SafetyGate, ChemGate, OptGate, LiveGate, ShotValidator, and more. Download and run locally.

9 NOTEBOOKS
TomoGate DemoLiveGate DemoSafetyGate DemoGuardGate DemoOptGate DemoChemGate DemoShotValidator Demo

FOR RESEARCHERS

Your First Experiment

1
Log in to your dashboard
2
Navigate to Research → Quick Run tab
3
Select "Compare Decoders" (recommended first experiment)
4
Click RUN NOW — results appear in the Results tab in ~2 minutes
5
View statistical analysis, export BibTeX, and download CSV data

Understanding Doom Detection

Doom rate is the probability your quantum job will produce unusable results. QubitBoost measures doom before and during execution, aborting hopeless jobs to save your budget.

<11%

QEC Viable

11-25%

Marginal

>25%

Abort zone

The doom threshold (default 25%) controls when jobs are aborted. Stricter thresholds (lower %) save more money but reject more jobs. Health classification uses Wilson score confidence intervals for statistical robustness with small sample sizes.

Using with Qiskit

Submit Qiskit circuits directly via the Python SDK:

from qubitboost_sdk import LiveSession
from qubitboost_sdk.applications.vqe import VQEDoomWrapper

# Doom detection: abort expensive circuits when hardware is unhealthy
with LiveSession(backend="ibm_fez", doom_threshold=0.7) as session:
    result = session.run(circuit, shots=4000)
    print(f"Doom rate: {result.doom_rate:.3f}")

# VQE-specific: sentinel probe before expensive optimization
wrapper = VQEDoomWrapper(doom_threshold=0.15, n_qubits=4)
vqe_result = wrapper.run_with_sentinel(
    parameters=[0.1, 0.2, 0.3], shots=10000
)

The SDK provides application-level wrappers (VQE, QAOA) and low-level session management. Install with extras: pip install qubitboost-sdk[qiskit] for IBM support.

Exporting for Papers

BibTeX

Research → Results → "Export BibTeX" generates a .bib file with experiment metadata and mandatory QubitBoost citation.

CSV Data

Export raw experiment data as CSV for analysis in Python, R, or other tools.

LaTeX Tables

Job detail pages include "Copy LaTeX" for publication-ready result tables.

Mandatory citation:

@software{qubitboost2026,
  title = {QubitBoost: Quantum Execution Governance Platform},
  author = {QubitBoost Team},
  year = {2026},
  url = {https://qubitboost.io},
}

Budget & Limits

FeatureAcademic Tier
Monthly Budget$500/month
BackendsIBM, Rigetti, IonQ, Google, IQM
Machine Activations2 concurrent
Research DashboardFull access
PublicationFree to publish (cite QubitBoost)

SUPPORTED BACKENDS

IBM Quantum

VALIDATED

Fez, Torino, Marrakesh (3 backends)

Rigetti

CALIBRATED

Ankaa-3 via AWS Braket

IonQ

CALIBRATED

Aria, Forte via AWS Braket

IQM

CALIBRATED

Garnet, Emerald

SDK DOCUMENTATION

Deep technical guides for each QubitBoost product. Full API reference, code examples, hardware integration patterns, and production deployment instructions.

Getting Started

Installation, setup, and your first session across all 7 products

TomoGate Guide

Pre-flight state certification, compressed tomography, fidelity thresholds

LiveGate Guide

Job lifecycle, sentinel probes, real-time abort, smart routing, 8 backends

SafetyGate Guide

QEC doom detection, 3 operating modes, adaptive thresholds, decoder integration

GuardGate Guide

QAOA solution validation, constraint checking, approximation ratio gating

OptGate Guide

Adaptive shot allocation, 128 median shots, multi-vendor (IBM/IonQ/IQM/Rigetti)

ChemGate Guide

ADAPT-VQE operator pruning, 32-42% savings, TrustGate safety, PySCF integration

ShotValidator Guide

Redundant syndrome verification, shot quality filtering, 100% error capture on IBM Fez

JUPYTER NOTEBOOKS

Interactive notebooks with working code examples. Download and run locally with pip install qubitboost-sdk[ml].

SafetyGate: QEC Doom Detection

Syndrome analysis, doom prediction, operating modes, V3 diagnostics

01_safety_gate_qec.ipynb

ChemGate: VQE Cost Reduction

Operator pool pruning, TrustGate safety, cost comparison

02_operator_selector_vqe.ipynb

OptGate: Shot-Efficient QAOA

Adaptive optimization, multi-vendor, cost savings analysis

03_adaptive_optimizer_qaoa.ipynb

LiveGate: Execution Gateway

Job submission, sentinel probes, real-time abort, health monitoring

04_livegate_execution.ipynb

Combined Pipeline

All gates working together — QEC, VQE, and QAOA end-to-end

05_combined_pipeline.ipynb

ShotValidator: QEC Shot Filtering

Redundant syndrome agreement, shot-level gating, hardware validation results

06_shot_validator_qec.ipynb

Digital Twin Calibrations

Load calibration data, extract noise metrics, select qubit chains, track drift

07_digital_twin_calibrations.ipynb

Adversarial QEC: Measurement Attacks

Corrupt measurement qubits, detect attacks with ShotValidator, evasion analysis

08_adversarial_qec_measurement_attacks.ipynb

QEC Resilience Benchmarking

Systematic resilience testing, attack sweeps, long-run monitoring, experiment framework

09_qec_resilience_benchmarking.ipynb

ENTERPRISE

Full Enterprise Documentation

Full enterprise documentation, production deployment details, and integration playbooks are available on request.

REQUEST ACCESS