Google Cloud API Gateway turns REST APIs into MCP tools

Software4 October 2026Updated: 16:051 min read

In brief

Google Cloud API Gateway now acts as a native remote Model Context Protocol (MCP) server. Developers can add a specific annotation to existing OpenAPI 3.x specs to turn standard REST operations into tools that AI agents can use. Agent traffic is then subject to existing authentication, quota, and logging policies.

What changed?

Google Cloud API Gateway now serves as a native remote MCP (Model Context Protocol) server. This means developers who want to expose their REST APIs to AI agents no longer need to build and maintain a separate middleware layer.

How it works

Developers add custom annotations such as x-google-api-management.mcp to their existing OpenAPI 3.x specifications. This turns standard REST operations into discoverable tools for agents. The gateway automatically transcodes incoming MCP JSON-RPC requests into REST calls.

Existing infrastructure is preserved

During transcoding, existing authentication, quota, and logging policies apply to agent traffic as well. In other words, current security and management rules take effect without requiring new infrastructure.

Why it matters

For teams looking to expose their REST APIs to AI agents, this update removes the burden of building and maintaining a custom middleware layer. Since existing security and quota policies are preserved, migration is low-effort.

Sources

—
Google Developers BlogPrimary source
Turn your REST APIs into MCP tools with Google Cloud API Gateway

#Google Cloud #MCP #REST API #AI agents #API Gateway