Quickstart Guide

Get the QECSync SDK installed and run your first surface-code syndrome decode in under 15 minutes.

Prerequisites

  • Python 3.10 or higher
  • QECSync SDK access credentials (request via the access form)
  • A calibration JSON file from your quantum hardware provider (or use the included synthetic noise generator)

Installation

Once you have received your access credentials, install QECSync from the private PyPI endpoint:

terminal 
$ pip install qecsync \
  --extra-index-url https://pkg.qecsync.com/simple/ \
  --index-url https://pypi.org/simple/
Successfully installed qecsync-0.9.2

Verify the installation:

terminal 
$ python -c "import qecsync; print(qecsync.__version__)"
0.9.2

Step 1 — Build a noise model

The noise model captures your device's per-qubit and per-gate error characteristics. You can load it from a calibration export or build a synthetic depolarizing model for testing:

quickstart.py 
from qecsync import NoiseModel

# Option A: from hardware calibration JSON
noise = NoiseModel.from_ibm_json("calibration.json")

# Option B: synthetic depolarizing (for testing)
noise = NoiseModel.depolarizing(p=0.003, n_qubits=49)

Step 2 — Initialize the decoder

Create a DecoderEngine configured for code distance 5 and your noise model:

quickstart.py 
from qecsync import DecoderEngine

eng = DecoderEngine(
    algorithm="mwpm",
    code_distance=5,
    noise_model=noise
)
print(eng.weight_matrix.shape)  # (49, 49)

Step 3 — Run a decode

Pass syndrome data (a 3D boolean array of shape [rounds, rows, cols]) to the decoder:

quickstart.py 
import numpy as np

# Generate synthetic syndrome: 10 rounds of d=5 surface code
syndromes = np.random.randint(0, 2, size=(10, 4, 4), dtype=bool)

result = eng.decode_rounds(syndromes, rounds=10)
print(result.logical_correction)  # PauliString("IIIII...")
print(result.latency_ms)          # 0.64

Next steps