Kong AI/MCP Gateway and Kong MCP Server Technical Breakdown
Blog post from Kong
The blog post delves into the Model Context Protocol (MCP) and its integration with Kong's AI Gateway to streamline the interaction between AI agents and enterprise systems. It highlights the role of MCP in managing data flow to prevent information overload in AI systems, allowing for standardized and efficient data processing. The Kong Gateway 3.12 release introduces several MCP capabilities, including the AI MCP Proxy and OAuth2 plugins, to facilitate secure and seamless communication between MCP clients and servers. These plugins enable the conversion of existing Kong Gateway services into MCP tools, providing real-time visibility, reducing vulnerabilities, and supporting advanced agentic patterns such as multi-step workflows and contextual enrichment. The article also discusses the architecture of MCP, detailing its client-server model and the transport mechanisms based on JSON-RPC 2.0, such as Stdio and Streamable HTTP. The integration of security measures like OAuth2 ensures that MCP tool endpoints remain protected, further enhancing the scalability and reliability of AI deployments. The post encourages exploring Kong’s flexible plugin ecosystem and its partnership with AWS to build secure and scalable MCP-enabled architectures.