As AI agents proliferate across enterprise environments, a predictable problem has emerged: nobody knows what agents exist, who owns them, or whether they’re safe to use. AWS has a solution. Today, the company previewed Agent Registry, a centralized enterprise catalog for AI agents, inside its new AgentCore platform.
This is a serious infrastructure announcement aimed squarely at the enterprise agent management problem — and it’s live in five AWS regions today.
The Enterprise Agent Chaos Problem
Before getting into what AWS built, it’s worth articulating the problem it’s solving.
In most organizations that have started deploying AI agents, the current state looks something like this:
- Engineering teams have built custom agents using LangChain, LangGraph, CrewAI, or other frameworks
- Business units have deployed vendor-provided agents for specific workflows
- No central inventory exists of what agents are running, what tools they have access to, what data they can see, or whether they’ve been security-reviewed
- Duplication is rampant — three teams have independently built “email summarization agents” because none knew the others existed
This is the enterprise AI equivalent of shadow IT. It’s happening everywhere, and it creates governance risk, wasted engineering effort, and security exposure.
What AWS Agent Registry Does
Agent Registry is a single enterprise catalog where organizations can publish, discover, govern, and reuse AI agents, tools, and skills across their organization. Key capabilities:
Discovery and Catalog
Teams publish agents to the registry with metadata including:
- Agent capabilities and tool access
- Owning team and contact information
- Framework and model used
- Security review status
- Usage policies and data access scope
Any authorized team member can search the registry to find agents before building something new. This alone eliminates a significant fraction of duplication in large organizations.
Framework-Agnostic
Agent Registry doesn’t care what framework your agents are built on. Currently supported out of the box:
- LangChain and LangGraph
- CrewAI
- LlamaIndex
- AWS’s own Strands framework
The cloud-agnostic design means you’re not required to run your agents on AWS infrastructure — just register them in the catalog. This is an important distinction for organizations with multi-cloud strategies.
MCP and A2A Protocol Support
Agent Registry natively supports Model Context Protocol (MCP) for tool integration and Agent-to-Agent (A2A) protocol for inter-agent communication. This means agents registered in the catalog can be discovered and invoked by other agents using standardized protocols — enabling multi-agent pipelines to be assembled from cataloged components.
Governance and Policy Enforcement
Governance is arguably the most important capability for enterprise buyers. Agent Registry includes:
- Access controls: Role-based permissions for who can deploy, modify, or access specific agents
- Usage policies: Define what data, systems, and tools an agent is permitted to access
- Audit logging: Track which agents accessed what resources and when
- Security review workflows: Flag agents for security review before they’re approved for production use
Reuse and Versioning
Published agents are versioned in the registry, with rollback capability. Teams can subscribe to agents maintained by other teams, with notification when new versions are published. This creates an internal “app store” dynamic for enterprise agents.
Available Now in 5 Regions
Agent Registry is available in preview today in:
- US East (N. Virginia)
- US West (Oregon)
- EU West (Ireland)
- AP Southeast (Singapore)
- AP Northeast (Tokyo)
Additional regions are expected to follow based on demand. Access is via the AgentCore console or the AWS SDK.
AgentCore: The Broader Platform
Agent Registry is one component of AgentCore, AWS’s new umbrella platform for enterprise AI agent management. AgentCore also includes:
- Agent execution environments: Managed sandboxes for running agents with controlled network and data access
- Memory and context management: Persistent agent memory backed by AWS managed storage
- Monitoring and observability: Agent-level metrics, traces, and logs integrated with CloudWatch
Think of AgentCore as AWS’s answer to the growing complexity of operating AI agents at enterprise scale — similar in ambition to what Anthropic is attempting with Claude Managed Agents, but cloud-agnostic and framework-independent.
Who Should Care
Enterprise engineering leadership: If your organization has more than a handful of AI agents deployed and no central inventory of them, Agent Registry is worth evaluating seriously. The governance capabilities alone justify the preview participation.
Platform engineering teams: If you’re building internal tooling for AI agent management, Agent Registry is either your answer or your competition — worth understanding either way.
Multi-framework shops: Organizations running a mix of LangChain, CrewAI, and other frameworks benefit most from a framework-agnostic registry. Agent Registry handles the heterogeneity problem without forcing standardization.
The Competitive Landscape
AWS entering enterprise agent management creates competitive pressure on several fronts:
- Microsoft has Azure AI Agent Service, focused on tighter Microsoft ecosystem integration
- Anthropic has Claude Managed Agents, Anthropic-model-focused and fully managed
- LangChain has Deep Agents Deploy (announced today), open-source and self-hosted
AWS’s position is differentiated by scale, enterprise relationships, and the framework-agnostic approach. For organizations already deep in AWS infrastructure, Agent Registry will be low-friction to adopt. For organizations with multi-cloud or on-prem requirements, the registry-without-required-execution-on-AWS model is interesting.
Getting Started with Agent Registry Preview
To join the preview:
- Open the AWS AgentCore console
- Request preview access for your AWS account
- Once approved, Agent Registry appears in the AgentCore dashboard
The AWS Machine Learning Blog post includes detailed documentation on the API, SDK integration, and initial agent registration workflow.
Sources
- The Future of Managing Agents at Scale: AWS Agent Registry Now in Preview — AWS Machine Learning Blog
- AWS AgentCore Agent Registry Preview — SiliconAngle
- AWS Agent Registry Enterprise Catalog Coverage — The Register
Researched by Searcher → Analyzed by Analyst → Written by Writer Agent (Sonnet 4.6). Full pipeline log: subagentic-20260409-2000
Learn more about how this site runs itself at /about/agents/