Quickstart

Install the SDK, generate your first synthetic dataset, and download an evidence bundle in under 5 minutes.

Prerequisites

A RadMah AI account (free tier works for this guide)
An API key from Settings → API Keys in the dashboard

ℹAPI key prefix

Live keys use the sl_live_ prefix. Test keys use sl_test_. Both work for this guide.

Install the SDK

Install

pip install radmah-sdk

Authenticate

Create a client instance with your API key. The SDK sends it via the X-API-Key header automatically.

Create client

from radmah_sdk import RadMahClient

client = RadMahClient(api_key="sl_live_your_key_here")

⚠Keep your key secret

Never commit API keys to source control. Use environment variables or a secrets manager in production.

Generate synthetic data

Describe the data you need in plain English. The AI Orchestrator translates your description into a sealed job specification (“sealed contract”), selects the best generation engine, and builds a plan automatically.

Create a job

job = client.jobs.create(
    description="customer orders for an e-commerce platform",
    rows=100_000,
)

print(f"Job {job.id} created — status: {job.status}")

✦No engine selection needed

You never need to pick a generation engine. The platform analyses your description and automatically selects the optimal engine, parameters, and statistical profile.

Approve the plan

Before generation begins, the platform produces a plan showing the schema, row count, and engine selection. Review and approve it programmatically or let it auto-approve.

Approve

# Retrieve the plan
plan = client.jobs.get_plan(job_id=job.id)

print(f"Engine:  {plan.engine}")
print(f"Columns: {[c.name for c in plan.columns]}")
print(f"Rows:    {plan.rows}")

# Approve to start generation
client.jobs.approve(job_id=job.id)

ℹAuto-approval

Pass auto_approve=True in the Python SDK create call, or include "auto_approve": true in the REST JSON body, to skip the approval step.

Wait for completion

Poll until the job finishes. The SDK handles polling automatically.

Wait

# Blocks until succeeded or failed (polls every 2s)
job.wait()

print(f"Final status: {job.status}")  # "succeeded"

Download results and evidence

Every job produces a CSV dataset and a cryptographically-chained signed evidence bundle. Download both.

Download

# Download CSV
job.download("output.csv")

# Or load directly into pandas
df = job.to_dataframe()
print(df.head())

# Download the full signed evidence bundle
client.artifacts.download_bundle(job_id=job.id, path="evidence/")

The signed evidence bundle

Core evidence sections shown below. Generated data artifacts are additional — the exact file count varies by generation and output shape.

#	Artifact	Purpose
1	Contract K	Machine-readable record of the approved job — schema, engine, seed
2	Run Manifest	Immutable execution record — job, tenant, engine, timestamps
3	Constraint Report	Per-column pass/fail for every declared constraint
4	Determinism Proof	Cryptographic hash proving reproducibility
5	Privacy Report	Differential privacy audit, k-anonymity, singling-out risk
6	Utility Metrics	Statistical fidelity metrics and ML utility score
7	Artifact Manifest	Index + hashes of the preceding artifacts — the tamper chain
8	Timing Telemetry	Stage-level wall-clock durations for the run
9	Cryptographic Seal	Binding hash over every prior artifact in the bundle

✦Evidence is always produced

Every job produces the full signed evidence bundle unconditionally, regardless of plan tier or dataset size. This cannot be disabled.

Next steps

Read Core Concepts to understand sealed job specifications, cryptographic seals, and signed evidence bundles
Set up Virtual SCADA for industrial simulation
Review the REST API Reference for direct HTTP integration