Govern your first AI action in 5 minutes
Free tier. No credit card. One SDK call sends an action through ARQERA's governance engine — AUTO, SOFT, or HARD — and returns a cryptographic evidence artifact.
Choose Your Path
Three ways to get started. Developer quickstart is first — skip ahead if you want personal or team setup.
For Developers
Build AI governance directly into your application. Python and TypeScript SDKs with full type safety, auto-retry, and cryptographic evidence trails.
For You
Set up Ore as your personal AI assistant. Morning briefings, email triage, calendar management, and more — all governed and auditable.
For Your Team
Deploy ARQERA for your organization with SSO, governance policies, approval flows, and a compliance dashboard ready for audit.
Developer Quickstart
Govern your first AI action in 5 steps
Sign up at arqera.io
Create your account in 30 seconds. Free tier starts immediately — no credit card required.
Get your API key
Navigate to Settings > API Keys > Create. Name your key and copy it — you'll only see it once. Store it in your environment variables.
# .env
ARQERA_API_KEY=ak_live_...
ARQERA_BASE_URL=https://api.arqera.io # optionalThe SDK reads ARQERA_API_KEY automatically if set.
Install the SDK
One command. Available for Python and TypeScript / Node.js.
Python
pip install arqeraTypeScript / Node.js
npm install @arqera/sdk
# or
yarn add @arqera/sdk
# or
pnpm add @arqera/sdkGovern your first AI action
Call client.governance.evaluate() with the action name, a description, and optional context. The governance engine checks all 7 laws and returns a verdict (PROCEED, ESCALATE, or BLOCK) in milliseconds.
1from arqera import ArqeraClient
2
3client = ArqeraClient(api_key="ak_your_key_here")
4
5# Evaluate an action against the 7 governance laws
6result = client.governance.evaluate(
7 "email.send",
8 description="Send quarterly report to investors",
9 context={"risk_level": "medium"}
10)
11
12print(f"Verdict: {result.verdict}") # "proceed" | "escalate" | "block"
13print(f"Explanation: {result.explanation}")
14print(f"Duration: {result.duration_ms}ms") # ~8ms
15
16# Each of the 7 laws reports individually
17for law in result.evaluations:
18 print(f" {law.law_name}: {law.result}")1import { ArqeraClient } from '@arqera/sdk';
2
3const client = new ArqeraClient({ apiKey: 'ak_your_key_here' });
4
5const result = await client.governance.evaluate({
6 action: 'email.send',
7 description: 'Send quarterly report to investors',
8 context: { risk_level: 'medium' },
9});
10
11console.log(result.verdict); // "proceed" | "escalate" | "block"
12console.log(result.explanation); // Why this verdict was reached
13console.log(result.duration_ms); // ~8msWhat the tier means
pending.email.send to external addresses is classified HARD by default. Your governance policies determine tier thresholds — adjust them in the Governance space.
Check the evidence trail
Every governed action produces a tamper-proof evidence artifact. List them by type, or export for EU AI Act and SOC 2 audit preparation.
1# List all governance evidence
2artifacts = client.evidence.list(artifact_type="governance_evaluation")
3for a in artifacts:
4 print(f"{a.action} → {a.verdict} ({a.duration_ms}ms)")
5
6# Export evidence for compliance audit
7export = client.evidence.export(
8 artifact_types=["governance_evaluation"],
9 start_date="2026-01-01",
10 end_date="2026-03-31",
11 export_format="json"
12)1// List all governance evidence
2const artifacts = await client.evidence.list({
3 artifact_type: 'governance_evaluation',
4});
5
6for(const a of artifacts) {
7 console.log(`${a.action} -> ${a.verdict} (${a.duration_ms}ms)`);
8}
9
10// Export evidence for compliance audit
11const exported = await client.evidence.export({
12 artifact_types: ['governance_evaluation'],
13 start_date: '2026-01-01',
14 end_date: '2026-03-31',
15 export_format: 'json',
16});You're done. Your first governed action is in the system. Open the Governance space to see it, or keep reading for environment variables and next steps.
Environment variables
| Variable | Description | Default |
|---|---|---|
ARQERA_API_KEY | Your API key — auto-read by the SDK if set | — |
ARQERA_BASE_URL | API base URL (change for on-prem deployments) | https://api.arqera.io |
ARQERA_TIMEOUT | Request timeout in seconds | 30 |
ARQERA_DEBUG | Enable verbose debug logging | false |
Personal Quick Start
Set up Ore as your personal AI assistant in 4 steps
Sign up at arqera.io
Create your account in 30 seconds. Your Early access available starts immediately — no credit card required.
Connect your accounts
Link Google, Microsoft, Slack, and Calendar via OAuth. One click each. Ore needs these to work on your behalf.
Meet Ore
Your AI friend who handles the small stuff. Ore learns your patterns, anticipates your needs, and keeps your day running smoothly.
Set your preferences
Decide what Ore can do autonomously (AUTO), what needs a heads-up (SOFT), and what requires your explicit approval (HARD).
What Ore handles for you
Ore learns your patterns over time and gets better at anticipating what you need. Everything it does is logged in your evidence trail so you always know what happened and why.
Team Quick Start
Deploy ARQERA for your organization in 6 steps
Create organization account
Sign up at arqera.io/auth and select the Team or Business plan. Name your organization and invite your first admin.
Configure SSO
Connect your identity provider — Okta, Microsoft Entra ID, or Google Workspace. SAML 2.0 and OIDC supported.
Set governance policies
Define your action tiers (AUTO / SOFT / HARD), approval flows, and compliance frameworks. The governance engine enforces them automatically.
Connect integrations
Link Slack, GitHub, Stripe, JIRA, and 100+ more tools via OAuth. One click each from the Connections space.
Deploy your first AI agent
Choose from the Agent Catalog or create a custom agent. Assign it a role, scope, and action tier. It starts working immediately.
Review evidence trail
Every AI action is logged with cryptographic evidence. Check the Governance space for your compliance dashboard and audit trail.
Supported identity providers
Okta
SAML 2.0
Microsoft Entra ID
SAML / OIDC
Google Workspace
OIDC
Custom IdP
SAML 2.0
Core Concepts
The building blocks of the ARQERA platform. Understand these and everything else clicks into place.
Action Tiers
AUTO — read and search freely. SOFT — act with a notification and 30-second undo. HARD — require explicit human approval before execution.
The Four AIs
Oju sees (observation and diagnostics). Ori thinks (reasoning and strategy). Ore acts (your personal AI friend). Ara speaks (the voice of the platform).
Evidence Chain
Every AI action produces a cryptographically signed artifact — who requested it, what the AI decided, why, and the outcome. Immutable and auditable.
Governance
Dual-brain evaluation (two independent AIs must agree) for high-risk actions. Disagreement escalates to human approval. No single point of failure.
App Spaces
The 12 spaces in the OS: Briefing, Agents, Canvas, Studio, Governance, Analytics, Connections, Graph, Inbox, System, Profile, and Help.
What's Next
Dive deeper into the platform.
API Reference
Full endpoint documentation with request/response schemas and interactive examples.
SDKs
Python and TypeScript client libraries with type-safe interfaces and auto-retry.
Governance Guide
Deep dive into action tiers, approval flows, dual-brain evaluation, and compliance.
Agent Catalog
Browse pre-built agents for investigation, operations, oversight, and custom builds.
Ready to govern your first AI action?
Request early access. No commitment required.