✦ Real IBM Quantum & OQTOPUS Hardware · Mac, iPhone & iPad

EGK Quantum
Circuits, Explained by On-Device AI

Write a circuit — or just describe one. Verify it locally, run it on real IBM Quantum or OQTOPUS hardware, then score the result against the ideal answer. When something breaks, an on-device AI reads the error and proposes a fix. Education, debugging and research in one workspace — on Mac, iPhone and iPad, with nothing uploaded.

🚀 Download on the App Store ▶ Watch It Run ⚛️ Learn Qiskit Free

Providers · IBM + OQTOPUS

100%

On-Device Reasoning

Generates, Fixes & Explains

Bytes Uploaded To Explain A Result

What EGK Quantum Does

Learn It. Debug It. Research It.

V27 grows EGK Quantum from a circuit runner into a full workspace in three chapters. Every job runs on real IBM Quantum Open Plan devices or OQTOPUS-hosted QPUs — never a classical approximation of the result.

Chapter 1

🎓 Education

Learn by running real circuits, not reading about them.

Curated lessons: Bell state, 3-qubit GHZ, Hadamard interference, phase kick
Ideal reference distribution beside every result
Live circuit diagram as you type
Plain-language AI explanation of every outcome

Chapter 2

🛠️ Debugging

Know your circuit is right before you spend QPU time.

Local OpenQASM 2 verification with clear errors for anything unsupported
Deterministic ideal simulation on your device
ISA / transpilation preview: “As Written” vs “As Submitted”
Fix with AI — reviewed and approved by you before it applies

Chapter 3

🔬 Research

Reproducible experiments with real numbers behind them.

Persistent experiment notebook
JSON & CSV export, raw counts preserved
Hardware-vs-ideal scoring and calibration snapshots
Cross-provider IBM vs OQTOPUS benchmarking

✨

Describe It, Get a Circuit NEW

Know the algorithm but not the syntax? Describe it in plain language and the on-device AI drafts OpenQASM for you — before any job is submitted.

✅

AI Suggestions, Verified NEW

AI-suggested optimizations are checked by a built-in statevector simulator to confirm the smaller circuit really computes the same thing. We verify the maths instead of trusting the model’s word.

🤖

AI Fixes Your Code

A circuit fails? Tap “Fix with AI.” The model reads the provider’s actual error and proposes a corrected version — nothing changes until you approve it.

🖥️

Real IBM & OQTOPUS Hardware

Submit to IBM Quantum Open Plan devices or OQTOPUS-hosted QPUs. OQTOPUS support follows the current job lifecycle, with automatic OpenQASM 2 → 3 conversion for supported circuits.

📏

Pre-Submit Workload Estimate NEW

See shots, mitigation mode, gate count and estimated QPU execution time before you press submit, so there are no surprises.

🗂️

Job History & Compare NEW

Every run is saved and searchable after relaunch. Select 2–4 runs and compare them side by side in a grouped probability chart, normalised so different shot counts stay comparable.

🔒

On-Device AI, Zero Uploads

Explanations, generated circuits and fixes come from a language model running locally. Your circuits and results never touch a third-party AI API.

🔑

Your Account, Your Keys

Use your own IBM Quantum API key and OQTOPUS credentials, stored only in the device Keychain. The app even tracks your key’s estimated expiry.

The Research Chapter NEW IN V27

Real Hardware, Real Metrics

Run once, then actually measure how close the QPU got. EGK Quantum compares every hardware result with a local ideal reference and keeps the complete record.

TVDClassical FidelityHellinger DistanceJensen-ShannonEntropyShot-Noise DiagnosticsParity & Z Expectation

📓

Experiment Notebook

Original QASM, IBM transpiled/ISA QASM, raw measurement counts, ideal distribution and metrics saved together, then exported as JSON or CSV.

📡

Calibration Snapshots

Captures per-qubit IBM calibration and gate/readout timing when available. Where a provider doesn’t publish a field, EGK Quantum shows nothing rather than inventing it.

🧭

Calibration-Aware Layouts

Ranks connected physical qubit regions on IBM backends using available calibration data, so you can place your circuit on the best qubits.

🎛️

Local Readout Mitigation

Corrects readout error as local post-processing from provider calibration data, with no additional QPU shots consumed.

⚖️

Cross-Provider Benchmark

Run the same circuit on IBM and OQTOPUS and compare the results head to head against the same ideal reference.

🖧

Platform-Aware Simulation

Ideal statevector reference up to 20 qubits on Mac and 16 on iPhone & iPad. Past the ceiling, the app says so instead of guessing.

The AI never does the maths. Every metric, ranking, mitigated distribution and workload figure is calculated deterministically in Swift. The on-device model only narrates the facts it is given, and its summaries are checked so they can’t introduce numbers that weren’t in the data. Zero-noise extrapolation is on the roadmap as the next mitigation stage — it is not part of this release.

Learn Qiskit NEW IN V27.3

Learn Qiskit. Run It Everywhere.

The same circuits, three ways to learn them: open-source lessons on GitHub, an interactive tool on the web, and the free nine-lesson Qiskit Track inside EGK Quantum V27.3 on iPhone and iPad.

On GitHub

💻 Qiskit Learning Lab

An independent, open-source teaching toolkit.

8 lessons, from qubits and gates to teleportation and the QFT
Runnable examples with exact statevector checks
MIT licensed, tested on Python 3.10–3.13
No hardware account needed

On the Web

🌐 Qiskit Learning Tool

Learn by watching the numbers settle.

Pick a circuit and read its Qiskit code
See the exact result next to sampled shots
Watch shot noise shrink as shots increase
Bell, GHZ, Hadamard interference and Grover

In the App · V27.3

📱 Qiskit Track 9 FREE LESSONS

Every lesson as OpenQASM and as native Qiskit Python.

Foundations: Bell, GHZ, interference, phase kick
Algorithms: Deutsch–Jozsa, Bernstein–Vazirani, Simon, Grover, phase estimation
Exact local simulation, no AI model, no network
Runnable script for your own IBM Quantum account

QiskitOpenQASMPythonIBM QuantumOQTOPUSOpen source

⚛️ Open the Qiskit Learning Tool GitHub Lab ↗ Get EGK Quantum

Qiskit is a trademark of IBM. EGK Quantum and Qiskit Learning Lab are independent projects, not affiliated with or endorsed by IBM. Simulation results are educational and are not evidence about real quantum hardware.

The Pipeline

From Idea To Research Result

Describe, verify, run, measure — one workflow, no manual polling and no separate analysis tools.

Write or Describe

Type OpenQASM or describe the circuit in plain language. The on-device AI drafts it, and the local verifier checks it before anything is submitted.

Estimate & Submit

Review the workload estimate, pick IBM Quantum or OQTOPUS, and submit. The circuit is transpiled for that exact backend automatically.

Real QPU, Real Counts

The job runs on actual hardware. Status is tracked for you and the raw measurement counts are preserved exactly as returned.

Score, Explain, Export

Results are scored against the ideal distribution, explained in plain language by on-device AI, saved to history, and exportable as JSON or CSV.

On-Device AI Models

Pick The Model That Fits

EGK Quantum ships with two on-device models, curated after side-by-side testing on real quantum job results. They power explanations, circuit generation, optimization suggestions and fixes. Switch between them anytime in Settings > AI Models.

Model	Best For	Size	Min RAM
Qwen 2.5 3B qwen2.5-3b-instruct-q4	Default Careful, complete result explanations	1.9 GB	8 GB
Phi-3.5 Mini phi-3.5-mini-instruct-q4	Deeper Reasoning Stronger step-by-step interpretation	2.4 GB	8 GB

Why only two: every model in the shared EGK catalogue was tested side-by-side against real Bell-state and GHZ-state jobs. Qwen 2.5 3B and Phi-3.5 Mini consistently gave the most careful, complete explanations — that list gets revisited only after a real re-test, not by default.

Why It Matters

Privacy By Architecture

Your quantum results are yours. EGK Quantum's reasoning layer never phones home to explain them.

🚫☁️

No Server Round-Trip

Result explanations and AI-suggested fixes are generated entirely on your device. There's no API endpoint receiving your circuit or your data.

🔬

Real Hardware, Real Results

Jobs run on actual IBM Quantum and OQTOPUS QPUs — the measurement counts you see are genuine quantum measurement outcomes, not a classical simulation.

🔑

Your Credentials Stay Yours

IBM and OQTOPUS credentials are stored only in the device Keychain, never in plaintext, never shared with EGK or any third party.

✋

AI Suggests, You Decide

"Fix with AI" proposes a corrected circuit for review — it never overwrites your work automatically. Every fix is one deliberate tap to apply.

From The Author · EGK Publishing

Companion Quantum Books

Go deeper with Isaac Khor’s technical and civilizational titles on hybrid AI systems, AGI, and the quantum divide — available as eBooks and on the EGK Book app with audiobook and extra features.

Subscription & Legal

Straightforward Terms

EGK Quantum offers a limited number of free circuit runs. Continued use beyond the free allowance requires EGK Quantum Pro, an auto-renewing monthly subscription. The exact subscription title, duration, and price for your region are always shown on the in-app purchase screen before you buy.

Payment is charged to your Apple ID account at confirmation of purchase. Subscriptions renew automatically unless auto-renew is turned off at least 24 hours before the end of the current period. Manage or cancel anytime in your Apple ID Account Settings. Restore Purchases is available on the in-app subscription screen.

By subscribing you agree to Apple's Standard End User License Agreement: apple.com/legal/internet-services/itunes/dev/stdeula

Read EGK Microelectronic Solutions Group Sdn. Bhd.'s full Privacy Policy: egkhor.com.my/privacy

EGK Quantum Circuits, Explained by On-Device AI

Every Single-Qubit Gate. One Wire.