Hello Apigee Community! ![]()
In our latest TechTalk, Tyler Ayers explores how developer portals are evolving from human-readable documentation catalogs into machine-discoverable neural hubs for agentic ecosystems. He demonstrates how to use Apigee and API Hub to deploy full-service, Model Context Protocol (MCP) powered AI portals for centralized agent access and corporate governance.
Watch the recording
Access the presentation
The agentic portal architecture 
To prevent tool sprawl and govern dynamic AI runtimes, your developer portal needs to act as a single source of truth for both developers and autonomous agents. Tyler highlighted three core architectural pillars:
-
Machine-discoverable cataloging: Transitioning from static OpenAPI docs to dynamic indexing using the Agentic Resource Discovery (ARD) specification and MCP registry formats.
-
Headless, serverless integration: Leveraging Apigee and API Hub APIs as a pure backend-for-frontend (BFF) engine to handle registration, subscription management, and metadata cataloging without needing a custom database.
-
Token & quota analytics: Providing clear visibility into token consumption for LLMs and call-based quotas for MCP tools directly through the Apigee Metrics API.
Key takeaways from the session 
Tyler provided a technical deep dive into building an out-of-the-box AI Portal solution.
He showcased how to deploy a lightweight proxy architecture that maps enterprise metadata stored in API Hub directly to the new Agentic Resource Discovery (ARD) format. By connecting an OIDC proxy mock with Apigee key-value maps (KVMs), developers can quickly test dynamic client registration (DCR), credential issuance, and user onboarding without external identity dependencies. Finally, Tyler demonstrated how agents can inspect MCP tool schemas (such as billing or customer lookup tools) directly within the self-service interface before calling backend proxies.
Leave Behind Resources 
-
Apigee AI Portals Solution Repository: github.com/gcp-samples/ai-portals-solution -
Agentic Resource Discovery (ARD) Announcement: developers.googleblog.com/announcing-the-agentic-resource-discovery-specification/ -
ARD Specification Specification on GitHub: github.com/ards-project/ard-spec
Next steps 
-
Continue the conversation: How is your team handling discoverability for your MCP servers and custom agents? Share your architectural ideas or ask Tyler your questions in the comments below! -
Share feedback: Help us shape future developer sessions by sharing your thoughts here → Feedback: Apigee Community TechTalks -
Explore further: Ready to architect your enterprise agentic strategy? Contact a Google Cloud sales specialist for guidance. -
Stay tuned: Keep an eye on the forum for updates on our upcoming sessions → Google Cloud Apigee Community TechTalks