Install, test, reproduce, evaluate.

QECTOR documentation is organized around one practical workflow: clone the real repository, build the Rust/Python extension, reproduce benchmark evidence, then decide whether the free non-commercial license or a commercial evaluation tier fits your use case.

Real public PowerShell path

git clone https://github.com/qectorlab/qector-decoder.git
cd qector-decoder
py -3.11 -m venv .venv
.\.venv\Scripts\python.exe -m pip install --upgrade pip maturin
$env:PYO3_PYTHON = (Resolve-Path .\.venv\Scripts\python.exe).Path
.\.venv\Scripts\python.exe -m maturin develop --release --no-default-features

The live repository does not currently ship install.py. The supported public path is a virtual environment plus maturin develop.

Most useful pages and repository files

Install

Setup and verification

Use the real repository build path first. Do not use old python install.py commands because that file is not present in the live repo.

Open install page
API

README and examples

The decoder repo README includes Python quickstart, examples, decoder stack, known limitations, and smoke-test context.

Open README
Benchmarks

Evidence and limits

Review LER parity, belief-matching evidence, BP-OSD/LDPC scope, GPU bit-identity, and what not to claim yet.

Open evidence
License

Free vs paid use

Personal, academic, educational, and non-commercial research use is permitted. Commercial use requires paid written permission.

Read license
Commercial

Why pay if PyMatching is free?

QECTOR is sold as a reproducibility and multi-decoder workflow platform, not as the fastest exact-MWPM baseline.

Read commercial fit
Workbench

Upcoming validation workstation

The planned differentiator for PDF benchmark reports, SHA-256 artifacts, environment snapshots, and local comparison workflows.

View roadmap

Commercial pilot checklist

  • Clone the canonical repo and build with maturin develop --release --no-default-features
  • Verify import with from qector_decoder_v3 import UnionFindDecoder
  • Run the benchmark and reproduction docs relevant to your workload
  • Compare against PyMatching/Stim for your actual circuits
  • Document saved engineering time, BP-OSD/LDPC needs, and reproducibility requirements
Request pilot
Claim hygiene.
Do not present QECTOR as faster than PyMatching or production-ready real-time hardware QEC infrastructure unless a specific reproducible artifact proves that exact claim. The current public moat is workflow, reproducibility, multi-decoder coverage, BP-OSD/LDPC, commercial legal clarity, and tested CPU/GPU bit identity.