The Core Update

Apigee hybrid version 1.17.0 is now live. This major update introduces Model Context Protocol (MCP) support. Previously, MCP ran on Google-managed infrastructure. Now, its data plane integrates directly within your Kubernetes cluster. Your APIs become agent-callable tools wherever your runtime resides.

Official Source: Google Announcement

Technical Impact & Mechanism

Teams adopted Apigee hybrid for critical reasons. Regulations, data proximity, or existing multi-cloud setups dictated local runtimes. These users needed AI agent access to their APIs. The challenge was enabling this without moving their operational infrastructure. This update addresses that directly.

Your core API interaction model remains. Define your API using an OpenAPI 3.0.x specification. From this, you create an MCP Discovery Proxy. Each API operation automatically becomes an AI agent tool.

Agents discover these tools via JSON-RPC tools/list calls. They initiate API operations using tools/call. Apigee handles the protocol translation and data transcoding.

No separate MCP server code is needed. You also don't operate a distinct MCP server.

The significant architectural shift is execution context. Your Kubernetes cluster now hosts the MCP data plane. The Apigee control plane, still in Google Cloud, generates the MCP tool configuration. Your cluster then fetches this configuration periodically.

Crucially, MCP request traffic stays local. Agent requests are received, authorized, and proxied within your infrastructure entirely. No MCP request or response crosses into Google Cloud during runtime.

Configuration updates are pull-based. A sidecar in your cluster retrieves new settings from Google Cloud. The Apigee control plane never initiates inbound network connections to your cluster. This means no inbound paths from Google are required through your firewall.

Here's how agents interact:

CONSOLE // JSON SYNTAX_CHECK: OK 
// Agent discovers available tools
{
  "jsonrpc": "2.0",
  "method": "tools/list",
  "id": 1
}

// Agent invokes a specific tool
{
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "tool": "your_api_operation_name",
    "arguments": { "param1": "value1", "param2": "value2" }
  },
  "id": 2
}

Action Plan for Developers & Businesses

  1. Upgrade Your Installation: Update to Apigee hybrid 1.17.0. Your cluster environment enables new MCP functionality.
  2. Standardize API Specs: Ensure all your APIs leverage accurate OpenAPI 3.0.x specifications. This powers tool generation.
  3. Configure Discovery Proxies: Set up MCP Discovery Proxies for your identified APIs. This exposes them to AI agents.
  4. Validate Network Policies: Confirm internal cluster traffic flow for agent-to-API communication. Ensure local security compliance.
Looking for deeper technical insights or tailored solutions? Check out my Case Studies & Work or Contact Waleed directly.
WR
AUTHOR // SOLUTIONS ARCHITECT

MUHAMMAD WALEED RAZA

I am a Solutions Architect and Digital Growth Engineer specializing in server-side tracking, CRM integrations, and conversion data recovery. I build deterministic, data-driven systems for high-growth brands and enterprise agencies worldwide.

HIRE ON UPWORK START A PROJECT