# Kong Catalog lets platform teams inventory and govern all MCP servers, agents and AI models in one place

Juhi Singh explains how organizations are struggling to manage MCP servers, AI agents, and AI models as these systems spread across teams faster than governance can keep up. The core problem is not just duplication or inconsistent protection, but a lack of visibility: many companies cannot say what they have, who owns it, which assets sit behind a gateway, or which dependencies could break when changes are made. Kong’s approach is to apply lessons learned from API management by extending familiar infrastructure to AI, including an AI gateway, registry, and especially Kong Catalog as a unified inventory and discovery layer. The catalog is positioned as a system of record for APIs, AI models, MCP servers, and agents, showing metadata, versions, documentation, ownership context, and whether assets are linked to the AI gateway for policy, observability, and control. Singh also highlights upcoming search, documentation, and versioning improvements, with agents currently in beta.