How to deploy an AI security platform while keeping data in your boundary.
The same security engine - the AI firewall, 40 detectors, agentic and MCP controls, output DLP egress gate, and evidence - runs on every tier. What changes across the three operating models is only placement and how your team operates: firewall only, firewall plus hosted control plane, and full private deployment. This page makes that boundary explicit so procurement and architecture reviews do not have to infer it from scattered copy.
Open Source
Self-hosted AI firewall plus guardrails.
The full security engine runs in your infrastructure
No control plane dependency to get protected
Local policy files and docs-based onboarding
Best for developer evaluation and small single-team rollouts
Growth
Self-hosted firewall with a hosted control plane.
The same security engine still runs in your infrastructure
Hosted control plane for policy workflow and operations
Metadata-first hosted visibility by default
Best for teams that need shared rollout and trace operations
Enterprise
Optional self-hosting for both firewall and control plane.
Private deployment for both firewall and control plane when required
Custom packaging, procurement, and deployment planning
Suitable for air-gap or security review constraints
Custom pricing and support model
The same security engine runs on every tier; only firewall and control plane placement change across Open Source, Growth, and Enterprise.
You get the same protection on every tier, so the choice is about operations, not security coverage. Start with Open Source if the question is still "does the self-hosted firewall fit our app?" Move to Growth when the question becomes "how do multiple people manage rollout and operations?" Move to Enterprise when the question becomes "where is the control plane allowed to run?"
Secure every prompt, agent, and tool call, in your boundary.
Pull the open-source runtime, drop it into your stack, and start enforcing policy in minutes, free, forever. Add the hosted control plane when you want centralized policies, evidence, traces, and team workflows.