Nanorix Python SDK

Sealed ephemeral capsules for regulated data, with independently verifiable AuditProofs.

Installation

pip install nanorix

Quick Start

import nanorix

client = nanorix.Client(api_key="nrx_live_...")

# Create a capsulefile (one-time setup)
capsulefile = client.capsulefiles.create(
    name="patient-risk-scorer",
    runtime="python3.11",
    entrypoint="python model.py",
    files=["model.py", "requirements.txt"],
)
client.capsulefiles.wait_until_finalized(capsulefile.id)

# Process data — capsule auto-destroyed after completion
outputs, cdp = client.capsules.run(
    capsulefile=capsulefile.id,
    inputs={"patient.csv": open("patient.csv", "rb")},
)

# The proof
print(cdp.verify_url)  # https://nanorix.io/verify/cap_...

Workspace Mode

For multi-step workflows with full control:

capsule = client.capsules.create(capsulefile="cf_...")

capsule.upload("patient.csv", data)
capsule.execute("python preprocess.py")
capsule.execute("python model.py")

prediction = capsule.download("prediction.json")
cdp = capsule.destroy()

# Share the proof with auditors
verification = cdp.verification()

AuditProof Access

Every destroyed capsule produces an AuditProof covering the full lifecycle — ingress, execution, egress, and destruction:

# AuditRecord — your private audit trail
full = cdp.full()
print(full.activity)       # Complete event timeline
print(full.destruction)    # 8-step chain verification

# AuditProof — safe to share
verify = cdp.verification()
print(verify.data_ingested)   # File count and hashes only
print(verify.residency)       # How long data existed
# Zero filenames, commands, or operational details

Error Handling

from nanorix import ExecutionError, TimeoutError

try:
    outputs, cdp = client.capsules.run(capsulefile="cf_...", inputs=inputs)
except ExecutionError as e:
    print(f"Exit code: {e.exit_code}")
    print(f"CDP still generated: {e.cdp}")  # Destruction happened
except TimeoutError as e:
    print(f"Capsule still alive: {e.capsule.id}")
    # Decide: retry, debug, or destroy

Async Support

import nanorix

async with nanorix.AsyncClient(api_key="nrx_live_...") as client:
    outputs, cdp = await client.capsules.run(
        capsulefile="cf_...",
        inputs={"data.csv": data},
    )

Configuration

client = nanorix.Client(
    api_key="nrx_live_...",
    base_url="https://api.nanorix.io",  # default
    timeout=30.0,                        # default
    max_retries=3,                       # default
)

Documentation

• Verify an AuditProof
• Console
• Verify an AuditProof