Deployment

Deploy HUMΛN agents in under 5 minutes with hosted, hybrid, or self-hosted options

Quick Start: 5-Minute Deployment

HUMΛN is designed for rapid deployment. Whether you're prototyping in minutes or deploying to production at scale, we support three deployment models:

Hosted
Zero-config deployment to HUMΛN Cloud
✅ Deploy in 2 minutes
✅ No infrastructure setup
✅ Auto-scaling included
Hybrid
Control plane in cloud, data plane on-prem
✅ Deploy in 10 minutes
✅ Data sovereignty
✅ Compliance-ready
Self-Hosted
Full control in your infrastructure
✅ Deploy in 12 minutes
✅ Complete control
✅ HIPAA/FedRAMP ready

Choose Your Deployment Model

>
SDK:

Container Deployment

Deploy agents as containers using Docker, Docker Compose, or Kubernetes. Portable across any cloud or on-premise infrastructure.

>
SDK:

Serverless Deployment

Deploy agents to serverless platforms for auto-scaling and pay-per-use pricing. Perfect for variable workloads.

>
SDK:

Environment Configuration

All deployment models use the same environment variables:

VariableRequiredDescription
HUMAN_API_KEYRequiredYour HUMΛN API key
HUMAN_BASE_URLOptionalAPI endpoint (defaults to hosted)
HUMAN_PASSPORT_IDRequiredAgent Passport ID
HUMAN_CAPABILITIESOptionalComma-separated capability list
NODE_ENVOptionalEnvironment (development, production)

Monitoring & Health Checks

Every deployed agent automatically exposes health and metrics endpoints:

GET /health200 OK

Basic health check (liveness probe)

GET /ready200 OK

Readiness check (dependencies available)

GET /metricsPrometheus

Prometheus-compatible metrics

Instant Visibility in Admin Console

The moment your agent deploys, it appears in the HumanOS Admin Console. No dashboards to build, no monitoring to configure.

Zero-DevOps Visibility

Deploy → Visibility. That's it. No Grafana to configure, no Datadog agents, no custom dashboards.

You Get Instantly:

  • Real-time agent health (healthy/degraded/critical)
  • Task completion rates & latency
  • Approval queues & escalations

No Configuration Required:

  • Auto-registered on first deploy
  • Provenance logging enabled
  • Compliance reports ready

Example: Deploy to Instant Visibility

$ npx @human/cli deploy

✅ Deployed: invoice-processor

📍 URL: https://api.human.ai/agents/inv-proc-123

🎛️ Admin Console: https://admin.human.cloud

🎉 Agent visible in Admin Console instantly:

  • • Status: Healthy
  • • Capabilities: finance/invoice/process
  • • Task queue: Empty (waiting for work)
  • • Provenance: Enabled, logging to ledger

# Or open directly from CLI:

$ npx @human/cli console

🌐 Opening Admin Console in browser...

What You See in Admin Console

Agent Overview
  • • Active instances
  • • Deployment timestamp
  • • Version history
  • • Health status
Live Metrics
  • • Tasks per hour
  • • Latency (P50/P95/P99)
  • • Error rates
  • • Escalation frequency
Audit & Compliance
  • • Provenance trail
  • • Approval history
  • • Boundary violations
  • • Compliance exports

Zero-Regret Migration

Start with hosted, migrate to self-hosted later with zero code changes. Only the HUMAN_BASE_URL changes.

Hosted → Hybrid

When you need data sovereignty:

  1. 1. Deploy data plane in your VPC
  2. 2. Configure secure tunnel to HUMΛN Cloud
  3. 3. Update HUMAN_BASE_URL
  4. 4. Done in minutes, not hours

Hybrid → Self-Hosted

When you need full control:

  1. 1. Deploy HumanOS services via Helm
  2. 2. Migrate control plane state
  3. 3. Point apps to new HUMAN_BASE_URL
  4. 4. Decommission hosted control plane

Ready to Deploy?