Overview
Most enterprises now run AI agents in many places at once: Amazon Bedrock AgentCore, SaaS platforms such as Agentforce and ServiceNow, data platforms such as Databricks, and coding assistants on developer laptops. Each comes with its own console and keys, so no one can say how many agents exist, who owns them, or what they can reach.
Lyzr Opencontroller is the control plane for that sprawl. It sits between your agents and everything they call, and keeps the inventory, identity, access, approvals and audit record for every agent, whichever cloud, framework or model it was built on.
What Opencontroller does:
-
Discover: read-only scans across clouds, clusters and developer laptops surface shadow agents and build one live inventory.
-
Register: one ownership record for every agent, across teams.
-
Identity: every agent gets its own identity.
-
Evals: every release is gated on evaluations.
-
Deploy: promote releases with built-in rollback.
-
Agent Gateway: control access to every agent.
-
LLM Gateway: route model calls and govern model spend by team and agent.
-
Guardrails: screen every input, output and tool call for prompt injection, PII and secrets.
-
Observability: OpenTelemetry traces tie every action to its owner and cost.
-
Refine: fix recurring failures from production data.
Works with your stack:
- Agent platforms: Amazon Bedrock AgentCore, Lyzr, Agentforce, ServiceNow, Databricks and Vertex AI.
- Identity: Okta, Microsoft Entra ID, SAML, OIDC and SCIM.
- CI/CD: GitHub, GitLab and Jenkins.
- Secrets: HashiCorp Vault and Azure Key Vault.
- Observability: OpenTelemetry
- Open standards: OpenAI-compatible APIs, MCP, A2A, REST, webhooks and CLI.
Deployment and security: OpenController deploys in your own AWS VPC, and air-gapped deployments are also supported. Control-plane data, credentials and telemetry never need to reach Lyzr. A first environment is typically live in less than a week.
Lyzr holds ISO/IEC 27001:2022 and ISO/IEC 42001:2023 certifications and a SOC 2 Type II report, and publishes HIPAA and GDPR compliance details in the Lyzr Trust Center. Lyzr is an AWS AI Competency Partner and a launch partner for the competency's Agentic AI categories.
How to buy: This solution is Deployed on AWS, so eligible purchases can count toward your AWS commitment. Choose Request private offer for pricing, or Request demo to see OpenController running on your own agents.
Use cases
AI for Dev and DevOps
Enterprise software teams face severe security, compliance, and sprawl challenges as AI-assisted coding tools and autonomous agents proliferate across pipelines. Implementing a centralized control plane provides complete visibility into AI agent behavior, model usage, and code-level data access. Operating directly inside your AWS VPC, it enables organizations to safely accelerate developer productivity while enforcing real-time guardrails and strict auditability.
Details
Introducing multi-product solutions
You can now purchase comprehensive solutions tailored to use cases and industries.
Products included
Features and programs
Financing for AWS Marketplace purchases
Pricing
Custom pricing options
Integration guide
Lyzr Opencontroller integrates non-invasively into your existing infrastructure by acting as a unified gateway between your AI agents, model runtimes, and downstream tools. It connects directly with your enterprise identity providers (such as Okta), CI/CD pipelines, secret managers, and observability platforms like OpenTelemetry and Datadog. Deployable directly inside your AWS VPC, hybrid, or air-gapped environment, Opencontroller secures and governs model calls and data access across multi-cloud runtimes including Amazon Bedrock AgentCore, Vertex AI, and Databricks without requiring changes to your underlying tech stack.