Q-Day Time Machine

01 Harvest now, decrypt later

The data you send today can be read tomorrow.

Attackers can record encrypted traffic now and decrypt it once a large quantum computer exists. Every company knows this. Almost none can say which systems break first, or when.

1,399

logical qubits to break RSA-2048, in under a week with fewer than a million noisy qubits.

Gidney, arXiv:2505.15917 (2025)

2,000

qubits in IBM Blue Jay, the fault-tolerant system on IBM's public roadmap for 2033 and beyond.

ibm.com/roadmaps/quantum

2035

is when NIST's draft transition plan disallows RSA and elliptic-curve cryptography.

NIST IR 8547, initial public draft (2024)

02 How it works

Scan, forecast, prove, migrate.

step 01

Scan

Reads live websites and source code in eight languages and exports a CycloneDX 1.6 cryptography bill of materials.

step 02

Forecast

Compares each algorithm's published cost in logical qubits and gates with IBM's roadmap, year by year.

step 03

Prove

Runs Shor's algorithm and Grover's search, built in Qiskit, on real IBM quantum computers, and steps through them gate by gate with our quantum-debugger library.

step 04

Migrate

Swaps in NIST's post-quantum standards, ML-KEM and ML-DSA, with a working hybrid handshake.

03 Real hardware

We ran the attacks on three IBM quantum computers.

Toy-sized keys, real machines: ibm_fez, ibm_kingston and ibm_marrakesh. Every pass rule was fixed in code before the runs, and two negative controls had to fail.

Shor factored 15 = 3 × 5 on ibm_marrakesh.

answers the algorithm should givehardware noise

4,000 shots, 173 two-qubit gates. The four ideal answers dominate, which reveals the period r = 4 and the factors.

Grover's search has a fingerprint noise cannot fake.

Chance of finding a random 3-bit key. Success climbs, peaks at the optimal 2 iterations, then falls when the search overshoots, as theory predicts.

45 / 45validation tests passed on 3 machines

24 / 24strict Shor tests passed on every machine and placement

14 / 14strict tests passed on ibm_kingston, repeat run included

10 / 18strict Grover tests passed

Every run beat random noise. Grover failed some strict checks because it must keep quantum phases intact through about 140 two-qubit gates, and on ibm_marrakesh the qubits it ran on decided the result: 22% on one set, 64% on another. That gap is why a real attack needs fault-tolerant machines.

04 First scan

We scanned our own website first.

It already uses post-quantum key exchange, so recorded traffic stays secret. Its certificate does not, so once signatures can be forged, the site can be impersonated.

quantyxio.cloud

2 of 4 at risk

Key exchange: X25519MLKEM768
Certificate: ECDSA P-256
Issuing CA: ECDSA P-384
Encryption: AES-256-GCM

05 Engineering

1,002 tests and a tool you can put in CI.

# fail the build on quantum-vulnerable crypto$ qday scan --path ./src --fail-on critical # when does RSA-2048 break?$ qday forecast RSA-2048 # export a CycloneDX 1.6 CBOM$ qday cbom --host example.com -o cbom.json # run Shor on IBM hardware$ qday shor --N 15 --a 7 --backend ibm_kingston

96% line coverage · lint with security rules · CI on Python 3.11 and 3.12 · non-root Docker image

Exhaustive quantum272
Post-quantum, CLI, dashboard, golden results196
Forecast and risk193
Security94
TLS handshakes, five certificate types72
Scanner rules, eight languages58

The tests caught the scanner missing classic key exchange in OpenSSL 3.5 output, the exact risk the tool exists to flag.
They blocked hostnames that OpenSSL would have read as command options.
They found five more defects, from a statistics edge case to a dashboard crash. All fixed.

2050 End of the line

Know your Q-Day before it knows you.

Scan a site or a codebase, see the year each algorithm falls, and get the post-quantum replacement.

Launch the app Fly back to 2026

When does your encryption die?