Amazon Web Services · 2026-04-09 · notable
AWS Agent Registry — governed catalog for AI agents and MCP servers
AWS launched Agent Registry (preview) inside Amazon Bedrock AgentCore — a private enterprise catalog with semantic search, approval workflows, and IAM governance for managing AI agents, tools, and MCP servers across organizations.
Amazon Bedrock AgentCore adds a private, governed catalog for discovering and approving AI agents and MCP servers across your organization.
What is it?
AWS Agent Registry is a new service inside Amazon Bedrock AgentCore, now in preview, that acts as a private catalog for all the AI agents, tools, MCP servers, and custom resources an organization deploys. It lets teams register, search, and govern agents in one place instead of relying on spreadsheets or ad-hoc docs. Early adopters include Zuora and Southwest Airlines.
How does it work?
The registry stores metadata for each resource and supports both manual registration via the console or API and automatic discovery from MCP and A2A endpoints. Search uses a hybrid of keyword matching and semantic similarity, so typing 'payment processing' surfaces agents tagged 'billing.' Approval workflows move records from draft through review to discoverable state. Access is controlled through IAM policies and OAuth (Custom JWT), with CloudTrail providing a full audit log. The service is available in five AWS regions and is also accessible as an MCP server, so agents can look up other agents at runtime.
Why does it matter?
Organizations deploying many AI agents run into the same problems: teams build duplicate agents without knowing one already exists, there is no review step before an agent goes live, and there is no audit trail when something goes wrong. Agent Registry gives enterprise platform teams the visibility and control layer that has been missing since agentic systems started shipping. The MCP-server interface is a particularly practical touch — it means agents can query the catalog at runtime without a custom integration.
Who is it for?
Enterprise AI/ML platform teams and CTOs managing large multi-agent deployments on AWS.