From fully air-gapped to cloud-native — Polygraf meets your infrastructure where it is, with no compromise on security or privacy.
Select a model to explore its architecture and best-fit scenarios.
Full control, your hardware
Deploy entirely within your own data center. No cloud dependency, no external connectivity required.
Desktop Overlay
Cloud flexibility, data sovereignty
Deploy on your own AWS, Azure, or GCP tenancy. All compute and storage remains in your VPC.
Enterprises on cloud-first infrastructure
Best of both worlds
Central governance in the cloud, SLM inference on-premises. Designed for distributed organizations.
Multi-location enterprises, global teams
Zero external connectivity
Completely isolated deployment. No internet, no cloud, no external calls. Maximum security posture.
Defense, intelligence, critical infrastructure
Every deployment model, every capability — at a glance.
All Polygraf components are shipped as Docker containers and fully orchestrated with Kubernetes. Deployment takes minutes, not days.
Full control, your hardware
One-command Kubernetes deploy
SAML 2.0, OIDC, legacy LDAP
Horizontal pod autoscaling
CPU
8 cores
RAM
32 GB
Storage
500 GB SSD
CPU
4 cores
RAM
16 GB
Storage
100 GB SSD
CPU
2 cores
RAM
4 GB
Storage
20 GB
CPU
16 cores
RAM
64 GB
Storage
200 GB SSD
CPU
8 cores
RAM
32 GB
Storage
200 GB SSD
Common questions from infrastructure and security teams.
For cloud and hybrid deployments, most organizations are fully operational within 1–3 days. On-premises deployments depend on your environment but typically take 1–2 weeks including integration and testing.
Yes. Polygraf is designed for portability. Your policies, configurations, and audit history migrate seamlessly between deployment models. Organizations typically start with cloud and transition to on-prem as their needs mature.
An air-gapped deployment means Polygraf has zero network connectivity to the internet. All updates are delivered via signed, offline packages. Ideal for environments that process classified or extremely sensitive data.
Yes. Private cloud and hybrid models support multi-region architectures across AWS, Azure, and GCP simultaneously. Each region runs its own SLM inference, with central governance optionally federated.
No. Polygraf operates in-line but includes a configurable fail-open/fail-close policy. In fail-open mode, requests pass through uninspected if the service is unavailable, maintaining business continuity.
Polygraf does not offer a multi-tenant SaaS option by design — all deployments are single-tenant to ensure data isolation. Private cloud is the closest equivalent and can be fully managed by Polygraf’s engineering team on your behalf.
Our solutions engineers will recommend the right deployment model for your environment
© 2026 Polygraf AI. All rights reserved.
Your download will start now.
Please provide information below and we will send you a link to download the white paper.