Cloud API Registry overview

Cloud API Registry lets you discover, govern, use, and monitor Model Context Protocol (MCP) servers and tools provided by Google, or by your organization through Apigee API hub.

Cloud API Registry supports your agentic AI development experience by helping you understand which APIs and tools you can use, along with their associated policies and restrictions. For example, when building an AI agent, you can use Cloud API Registry to find, enable, and monitor the MCP servers and tools that your agent needs.

Key benefits

Cloud API Registry lets you better utilize MCP resources in your AI applications by providing you with the following benefits:

  • Unified discovery—to find all available MCP servers and tools in one place.
  • Centralized governance—to manage and control access to MCP servers and tools across your organization.
  • Simplified integration—to discover and understand capabilities and connection details.
  • Broad support—to integrate metadata from services provided by Google, or by your own organization through Apigee API hub.

Key concepts

To effectively use Cloud API Registry, it's important to understand the following concepts:

  • The Model Context Protocol (MCP) is an open, standardized protocol for connecting AI applications to external systems. You can extend the capabilities of your AI applications and agents by using MCP to connect to the data sources, tools, and services provided by MCP servers. Learn more about the MCP and how it works.

  • An MCP server is an endpoint that exposes functionalities known as MCP tools. You can host these servers on Google Cloud, your own infrastructure, or through third parties. MCP servers allow MCP tools to be discoverable and usable by AI agents and applications. Learn more about Google Cloud MCP servers.

  • An MCP tool is a specific function or capability provided by an MCP server that an AI agent or application can invoke. Each tool has defined schemas for its inputs and outputs. You must first enable an MCP server before an MCP client can access its tools.

  • Apigee API hub is a platform that helps you manage your organization's APIs, including MCP servers and tools. It serves as a central repository for API and MCP information. Cloud API Registry imports metadata from Apigee API hub, making MCP resources easier for AI agents to discover and use. Permissions through an Apigee API hub role are required to list MCP servers and tools using Cloud API Registry. Learn more about Apigee API hub.

  • Service Usage is a service that lets you control enabling and disabling MCP endpoints across your organization's resource hierarchy: projects, folders, and organizations. Permissions through a Service Usage role are required to enable and disable MCP servers using Cloud API Registry. Learn more about Service Usage.

Primary roles and uses

Cloud API Registry is designed for two primary roles:

  • Platform administrators who are responsible for managing the registry and who can enable and disable MCP servers for a project. After an MCP server is enabled, an AI agent can use its MCP tools. For more information, see Enable and disable MCP servers.

  • Application developers who use the registry to find and understand which MCP tools to use when building AI applications and agents. For more information, see Discover and list MCP tools.

What's next