Quantum computing progress, with the receipts.
How fast are quantum computers really improving, and how far are they from being useful? Provenova answers in plain English from public evidence: published experiments, real calibration records from cloud machines and independent benchmarks, each labelled by how much it can be trusted.
Running your own experiments? Make your results verifiable with the same open provenance format.
Is it getting better, and how far is it from useful? Long-run trends from public data, explained.
Every machine in the record, with its dated history, independent benchmarks and how it compares.
Bind your own runs to the exact hardware state that produced them, and publish results anyone can verify.
Make your own results verifiable
The tracker can only be as good as the evidence behind it. Provenova's open-source SDK binds every run to the exact calibration and hardware state that produced it, so a result is reproducible, comparable, citable and auditable across IBM, IonQ, Braket, Azure and simulators. There are no paid plans, and it works offline.
pip install provenova
From a run to an audit-ready result
Four steps, all built on one immutable, hash-chained record.
Wrap your job with @ql.capture. The circuit, backend, calibration and result are recorded locally — offline, no account.
Re-run against a drifted device state and score how much changed, with a Hellinger fidelity and a plain-language verdict.
Turn any run into a public, citable Result Card with a permanent persistent identifier — plus an optional DOI via Zenodo — provenance hash and embeddable badges.
Map runs to standards like FAIR or IEEE P7131 and issue signed, publicly-verifiable attestations.
Everything a result needs to be trusted
A Merkle run-hash binds circuit, backend, calibration and result into a tamper-evident, hash-chained ledger.
Deterministic drift profiles + Hellinger / TVD scoring show exactly why a result changed.
Public cards with a persistent identifier — plus an optional DOI via Zenodo — citations (BibTeX/RIS) and shields.io-style badges.
Evidence auto-collected from your runs; issue Ed25519-signed, revocable attestations.
A dated, public calibration and benchmark history behind the tracker and leaderboard.
One @ql.capture decorator; pluggable connectors for IBM, IonQ, Braket, Azure & simulators.
Built for how you work
Make results reproducible and citable. Publish a Result Card with a citable identifier and prove it hasn't changed.
Create an account →Compliance frameworks, signed attestations and a public Trust Center to share your posture, all included.
Create an account →Run the whole server on your own infrastructure, production included, under its source-available license.
Self-hosting guide →One decorator. Any backend. Fully offline.
Capture runs from your existing code with no rewrites, inspect them with the ql
CLI, and push to a hosted or self-hosted server when you're ready. The provenance format is open and
verifies its own hash without us.
import provenova as ql
from qiskit_aer import AerSimulator
@ql.capture(project="bell")
def run():
return AerSimulator().run(qc, shots=4096)
# $ ql list
# $ ql reproduce <id> --days 90 --profile bad_day
# $ ql push
Make your next quantum result reproducible.
No account needed to try it locally.