Skip to main content
15 backend families powering 17 decoder configurations · GPU Batch · Stim · Sinter · Qiskit

User Manual

Complete guide to QECTOR Decoder v3: installation, decoder selection, local measurements, diagnostics, and ecosystem integration.

Overview

QECTOR Decoder v3 is a Python library for quantum error correction (QEC) decoding. v1.0.1 is the current stable release (v1.0.0, August 2026, was the first): it provides 17 decoder configurations and helpers integrated into a high-performance compiled Rust core with a plug-and-play Python API.

Ecosystem Native: Swappable backend design. You can easily plug QECTOR into Stim, PyMatching, Sinter, and Qiskit pipelines with minimal API modifications: Sinter decoders plug in with custom_decoders=qector_sinter_decoders(), and a qiskit-qec plugin is registered.

Key Highlights

v1.0.0: First Stable Release

API stability tiers, Relay-BP, CS-OSD(lambda, w), colour-code cluster_bposd, qector CLI + qector-doctor, Sinter/qiskit entry points, pymatching shim.

Validate on Your Hardware

Hardware, GPU, and measurement results are device-local. Use the package diagnostic and declared measurement workflow instead of treating site text as performance evidence.

Multi-Algorithm Diversity

17 decoder configurations from exact Blossom MWPM to Belief-Matching, BP-OSD, and GPU batch decoding.

Compiled Rust Core

A compiled Rust core with a Python API. Exact memory, latency, and throughput behavior depends on the selected workload and environment.

Optional GPU Paths

GPU paths are optional and environment-dependent. Read the installed release diagnostics before relying on them.