For many teams, the decision to run AI agents in production hinges on where the control layer lives.
Yes, Lyzr's control plane can be deployed privately inside your own environment.
This is a core design principle of OpenController by Lyzr, the control plane for enterprise AI agents. It’s built to be cloud-agnostic, giving you the choice between our managed cloud service and a self-hosted instance in your VPC or on-premise data center.
A control plane is the neutral layer that governs everything beneath it. For AI agents, OpenController is that layer, and it's designed to run wherever your agents run.
Why Private Deployment Matters for Agent Governance
Choosing a deployment model is an architectural decision with security implications.
For organizations in regulated industries or with strict data policies, a private deployment is not just a preference. It's a requirement.
Data Sovereignty: Your agent's operational data, including traces and logs, never leaves your network boundary. This is critical for meeting data residency rules under GDPR and other regional mandates.
Security Posture: Self-hosting minimizes the external attack surface. You control the network policies, ingress/egress points, and security configurations, aligning the control plane with your existing security posture.
Latency and Performance: Co-locating the control plane with your agents and private data sources can reduce network latency. This ensures faster policy enforcement and observability lookups.
How Self-Hosted Deployment Works
We provide OpenController as a set of containerized services.
This package is designed for deployment into your own Kubernetes cluster, whether it’s running in a public cloud VPC (like AWS, Azure, or GCP) or on-premise.
You manage the underlying infrastructure. We provide the application.
This means you have full control over:
- The compute resources allocated to the control plane.
- The networking rules that govern it.
- The storage and persistence of all operational metadata.
The observability data that OpenController generates, like end-to-end traces of multi-agent calls, is written to databases that you own and manage. This creates the verifiable audit trail needed to prove compliance for agents handling PII and PHI.
Comparing Deployment Models
The right choice depends on your team's operational capacity and security requirements.
| Capability | Lyzr Cloud (Managed) | Private Deployment (Self-Hosted) |
|---|---|---|
| Infrastructure | Managed entirely by Lyzr. | Managed by your team in your environment. |
| Data Residency | Data is stored in Lyzr's secure cloud. | All data resides within your network boundary. |
| Network Control | Standard secure SaaS access. | Full control over all network policies. |
| Maintenance | Updates and patches are applied by Lyzr. | Your team manages the update schedule. |
| Initial Setup | Near-instant. Sign up and connect your agents. | Requires deployment to your infrastructure. |
Both models provide the exact same governance and observability features. The only difference is where it runs.
Conclusion: Your Environment, Your Control
The ability to self-host OpenController isn't an afterthought. It's central to our philosophy of giving enterprise teams the control they need to run AI safely.
You shouldn't have to compromise your security architecture to govern your agents. With a private deployment, you own the infrastructure, the network, and the data.
The core principle is simple: you must be able to control your AI infrastructure. To discuss how a private deployment would fit your architecture, let's talk.
Want to see how it fits your stack? Talk to us about OpenController by Lyzr.
Got a use case in mind?
8 weeks from use case to agents running in production.
Platform, people and FDEs, all in. Bring your environment. We'll co-build and stay until it's live.
Talk to UsFree resource
101 AI Agents Use Cases
100+ production-ready AI agent use cases across Marketing, Sales, HR, Banking, Insurance, Operations and a lot more
Download now
