Getting Started with AtlasAI
New customer? Start with Subscribe & get access — sign up, receive your tenant URL, first login, and in-product onboarding.
AtlasAI supports three deployment models. Choose the one that best fits your organization’s requirements around data residency, compliance, and operational control.
Deployment Models at a Glance
| Model | Best For | Time to First Value | Infrastructure Required |
|---|---|---|---|
| Shared SaaS | Teams wanting zero ops overhead | ~5 minutes | None |
| Dedicated TP | Isolated TP + DB per customer (managed by Atlas) | Varies (sales-led) | None (Atlas-hosted) |
| On-Prem | Organizations with strict data residency | ~30 minutes | Docker host (4+ CPU, 8 GB RAM) |
| Hybrid | Enterprises needing data-local processing with cloud AI | ~45 minutes | Docker host + network to Atlas Cloud |
For onboarding, provisioning, and database responsibilities across all modes, see Tenant Plane deployment models.
Dedicated Tenant Plane
A separate Tenant Plane and database for your organization, still integrated with Atlas Cloud Control Plane. Ideal when you need infrastructure isolation without running TP yourself.
Shared SaaS
The fastest path to production. Your tenant runs in Atlas Cloud alongside the Control Plane. No infrastructure to manage, automatic upgrades, and instant access to all modules.
Prerequisites: A modern browser and an email address.
Get started with Shared SaaS →
On-Prem Installation
Deploy the Tenant Plane inside your own infrastructure. All operational data — logs, metrics, traces, incidents — stays within your network boundary. The Control Plane remains in Atlas Cloud for tenant provisioning and AI model routing, but no raw telemetry leaves your environment.
Prerequisites:
- Docker Engine 24+ or Podman 4+
- Host with 4+ vCPUs and 8 GB RAM (16 GB recommended)
- Outbound HTTPS (443) to the Control Plane (set via
NEXT_PUBLIC_CP_URLat build; defaultcp.atlastechlab.com) - DNS record or static IP for the Tenant Plane UI
Hybrid Deployment
Combine the data locality of on-prem with the managed intelligence of Atlas Cloud. The Tenant Plane runs in your infrastructure while the Control Plane handles AI inference, model management, and cross-tenant analytics.
Prerequisites:
- Everything required for On-Prem, plus
- Network path from Tenant Plane to Control Plane on port 443 (default
cp.atlastechlab.com) - Edge Agent installer access (one per monitored host)
Next Steps
After deploying AtlasAI, follow the Your First Incident tutorial to create an incident, run root cause analysis, and generate your first automated runbook.