Run QCOS and QuantumLock with your platform compute

Customers ask four different things when they say “run QCOS on my compute”. This page answers each one honestly.

1. Every quantum job already runs on QCOS

When you submit a QUANTUM_CIRCUIT workload — from a quantum instance or straight to the Execution Plane — the platform dispatches it to the hosted QCOS runtime at api.softquantus.com, which routes it to simulators or real QPUs (IBM, IonQ, Quantinuum, Rigetti, Pasqal). You consume QCOS as a managed service; there is nothing to install or operate.

2. Free local simulation on your own machine

The softqcos SDK can execute aer* simulator backends locally with Qiskit Aer — on your laptop, your CI, or any machine you already own. Local simulation consumes no API credits and the circuit never leaves your machine; it is the recommended inner loop before you spend on hosted targets.

from softqcos_sdk import QCOSClient

client = QCOSClient()
result = client.simulate(circuit, shots=1000)  # runs locally on Qiskit Aer

3. Your containers on purchased GPU/CPU can use both products

GPU & CPU compute runs any container image and command you supply. Inside that container you can call the QCOS runtime and QuantumLock exactly as you would from anywhere else — pass your key in as an environment variable through your image or command:

# inside your GPU job's command
export QCOS_API_URL="https://api.softquantus.com"
export QCOS_API_KEY="sq-live-..."
python hybrid_pipeline.py   # classical ML on the GPU + circuits via QCOS

This is the supported pattern for hybrid pipelines: heavy classical stages on the GPU you bought, quantum stages on the hosted runtime, one wallet and one evidence trail across both.

4. Running the products on hardware you own

The paths below are licensed, sales-assisted deliveries — they exist as license machinery today, not as self-service downloads. Talk to SoftQuantus if you need one.

QCOS on your own machines (OEM / on-premises). QCOS supports per-machine OEM licenses and enterprise platform licenses, backed by an on-premise license server and per-device HAL entitlements. Distribution of the QCOS runtime binary is part of a commercial agreement; there is no public runtime image to pull. See the OEM & hardware integration notes in the QCOS section. QuantumLock on your side. What is public and supported today:

The client — pip install quantumlock (client-only wheel, with pqc, cli and quantum extras) talking to the hosted service at quantumlock.softquantus.com. See QuantumLock installation.
License activation binds a license to a machine fingerprint via POST /api/v1/licenses/activate and is verified with POST /api/v1/licenses/validate; a second machine is rejected. See Activation.
The QuantumLock server is not a customer artefact — there is no public server image or self-hosting path. Enterprise offline packages are negotiated through sales, as described in Deployment.

Quick decision table

You want	Do this	Cost model
Run circuits without operating anything	Submit to the Execution Plane / a quantum instance	Credits (simulators free)
Iterate on circuits for free	`softqcos` SDK local `aer*` simulation	Free, your machine
Hybrid classical+quantum pipeline	Your container on GPU/CPU compute calling QCOS from inside	Per-minute GPU + per-shot QPU
QCOS runtime on hardware you own	OEM / enterprise license via sales	Commercial agreement
QuantumLock in your product	`pip install quantumlock` against the hosted service	QuantumLock pricing

Run compute

QCOS & QuantumLock on compute

Guarantees

Run QCOS and QuantumLock with your platform compute

1. Every quantum job already runs on QCOS

2. Free local simulation on your own machine

3. Your containers on purchased GPU/CPU can use both products

4. Running the products on hardware you own

Quick decision table

​1. Every quantum job already runs on QCOS

​2. Free local simulation on your own machine

​3. Your containers on purchased GPU/CPU can use both products

​4. Running the products on hardware you own

​Quick decision table

1. Every quantum job already runs on QCOS

2. Free local simulation on your own machine

3. Your containers on purchased GPU/CPU can use both products

4. Running the products on hardware you own

Quick decision table