---
type: "post"
id: "announcing-mcp-support-in-apigee-hybrid-agentic-tools-2e5d91"
slug: "announcing-mcp-support-in-apigee-hybrid-agentic-tools-2e5d91"
title: "Announcing MCP support in Apigee hybrid: agentic tools that run in your own cluster"
description: "Apigee hybrid’s MCP support changes where an agent-facing API adapter executes, not the underlying API-to-tool model."
retrieval_nugget: "Apigee hybrid’s MCP support changes where an agent-facing API adapter executes, not the underlying API-to-tool model."
published_at: "2026-10-15"
updated_at: "2026-10-15"
record_date: "2026-10-15"
date_kind: "published_at"
topics: ["mcp"]
source_urls: ["https://developers.googleblog.com/announcing-mcp-support-in-apigee-hybrid-agentic-tools-that-run-in-your-own-cluster"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "post:announcing-mcp-support-in-apigee-hybrid-agentic-tools-2e5d91"
status: "published"
visuals: []
editorial_provenance: {"schema_version":"newruntime-editorial-copy-v1","content_status":"source_grounded_final","final_copy_sha256":"sha256:f961a79f4d34917479f2865076dbc69be243ee3d9369099fa978ce734626ea22","reviewed_at":"2026-10-15T04:04:59.203Z","source_evidence_count":1,"verified_claim_count":3,"site_analysis_schema_version":"newruntime-site-analysis-v1","site_object_kind":"field_note","observed_fact_count":3,"implication_count":2,"watch_condition_count":1,"related_record_count":0}
analysis: {"schema_version":"newruntime-site-analysis-v1","object_kind":"field_note","thesis":"Deployment location is part of the agent integration contract: teams can expose APIs as MCP tools without moving request execution out of the infrastructure where those APIs already run.","observed_facts":[{"text":"Apigee hybrid 1.17.0, released September 14, 2026, adds MCP support with the data plane running inside the customer's Kubernetes cluster.","source_urls":["https://developers.googleblog.com/announcing-mcp-support-in-apigee-hybrid-agentic-tools-that-run-in-your-own-cluster"]},{"text":"An OpenAPI 3.0.x specification can generate an MCP Discovery Proxy whose operations are exposed as tools through JSON-RPC discovery and calls.","source_urls":["https://developers.googleblog.com/announcing-mcp-support-in-apigee-hybrid-agentic-tools-that-run-in-your-own-cluster"]},{"text":"The announcement says request and response traffic stays within the cluster while a sidecar pulls configuration from the Apigee control plane.","source_urls":["https://developers.googleblog.com/announcing-mcp-support-in-apigee-hybrid-agentic-tools-that-run-in-your-own-cluster"]}],"mechanism":"The cluster-side data plane serves MCP requests locally; a sidecar fetches tool configuration from the control plane, which does not initiate an inbound connection into the cluster. Existing Apigee proxy controls, including API products, apply.","why_now":"The release extends Apigee's existing MCP approach to hybrid customers whose runtime is intentionally on-premises, in another cloud, or in a regulated environment.","implications":["Platform teams can evaluate agent access to existing APIs without treating migration to Google-managed runtime as a prerequisite.","Teams can apply existing Apigee governance to agent-facing tools, subject to their deployment and authorization configuration."],"evidence_boundary":"The vendor announcement describes architecture and setup, but supplies no independent deployment results, latency comparisons, or customer outcome measurements.","watch_conditions":["Whether operators can enable and govern this optional feature reliably across real hybrid deployments, including its staged canary rollout and configuration-fetch behavior."],"related_records":[],"new_branch_reason":"No prior New Runtime record was supplied for this specific hybrid MCP release; this creates a distinct record about its deployment boundary and control/data path."}
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/posts/announcing-mcp-support-in-apigee-hybrid-agentic-tools-2e5d91/","markdown":"https://newruntime.com/posts/announcing-mcp-support-in-apigee-hybrid-agentic-tools-2e5d91.md","json":"https://newruntime.com/posts/announcing-mcp-support-in-apigee-hybrid-agentic-tools-2e5d91.json"}
related_patterns: ["agent-protocols-become-interoperability-layer"]
---

# Announcing MCP support in Apigee hybrid: agentic tools that run in your own cluster

## Retrieval answer

Apigee hybrid’s MCP support changes where an agent-facing API adapter executes, not the underlying API-to-tool model.

Apigee hybrid’s MCP support changes where an agent-facing API adapter executes, not the underlying API-to-tool model. Google says Apigee hybrid 1.17.0, released September 14, 2026, runs the MCP data plane inside the customer’s Kubernetes cluster. An OpenAPI 3.0.x specification can generate an MCP Discovery Proxy whose operations are exposed as tools through JSON-RPC discovery and calls.

Thesis: Deployment location is part of the agent integration contract: teams can expose APIs as MCP tools without moving request execution out of the infrastructure where those APIs already run. Mechanism: The cluster-side data plane serves MCP requests locally; a sidecar fetches tool configuration from the control plane, which does not initiate an inbound connection into the cluster. Existing Apigee proxy controls, including API products, apply. Why now: The release extends Apigee's existing MCP approach to hybrid customers whose runtime is intentionally on-premises, in another cloud, or in a regulated environment.

First implication: Platform teams can evaluate agent access to existing APIs without treating migration to Google-managed runtime as a prerequisite. First watch condition: Whether operators can enable and govern this optional feature reliably across real hybrid deployments, including its staged canary rollout and configuration-fetch behavior. Evidence boundary: The vendor announcement describes architecture and setup, but supplies no independent deployment results, latency comparisons, or customer outcome measurements. The release supports the agent-protocols-become-interoperability-layer pattern while leaving operational performance and adoption unproven.

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- topic hub: [Mcp](https://newruntime.com/topics/mcp.md): Explore the mcp topic hub.
- collection: [Field Notes](https://newruntime.com/posts.json): Open the latest editorial analysis.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)