- Kong AI Gateway to Control LLM, MCP, and Agent Calls… Same Authentication, Rate Limiting, and Audit Logging Applied
- Expands AI Governance Portfolio Through Deployment, Integration, and Technical Support Tailored to Customer Environments
CloudNetworks (CEO Heung-gyun Jung) announced that it has signed a partnership with Kong, a global API and AI connectivity company, and will begin supplying Kong AI Gateway domestically. Kong AI Gateway is an AI gateway that centrally manages an enterprise's AI traffic through a single gateway. This partnership was pursued to support Korean enterprises in building AI operations frameworks, amid growing demand for control and governance over AI calls as generative AI and LLM adoption spreads into actual business operations.
As enterprise AI adoption moves from the experimentation stage into service operations, how to control and observe AI calls is emerging as a new operational challenge. Because LLM calls are also made in API form, the need for a central gateway capable of authentication, security, and monitoring is growing. In addition, as calls between multiple LLM providers, RAG, MCP, and agents become intertwined, it is becoming increasingly difficult to consistently manage AI traffic with existing API gateways alone.
Kong is a global company that develops API and AI connectivity technology, and has provided an API management platform based on Kong Gateway, which is widely used as open source. Recently, Kong has been expanding its management scope in the API management space to include LLM, MCP, and agent-to-agent (A2A) traffic, with Kong AI Gateway responsible for managing AI traffic within this scope.
Kong AI Gateway is designed to extend the existing API gateway concept to centrally control, secure, and observe LLM and generative AI API calls. It runs on the same runtime as Kong Gateway and manages calls from multiple LLM providers through a single standardized interface.
In terms of key features, Kong AI Gateway applies the same API security policies—such as authentication/authorization, rate limiting, and audit logging—to AI calls as well. It also provides AI-specific functions at the gateway layer, including PII removal, prompt control, semantic caching, and traffic observability, enabling security and policy measures that were previously implemented individually per application to be standardized into a common operational layer.
As AI adoption moves further into the service operations stage, enterprises are increasingly required to design not only for model performance but also for control over call pathways, cost management, and audit systems. Along with supplying Kong AI Gateway domestically, CloudNetworks will provide architecture design, deployment, integration, and technical support tailored to customer environments. Through this, the company plans to connect gateway-layer control systems with existing AI operations environments and expand its AI governance portfolio.
CloudNetworks CEO Heung-gyun Jung said, "As generative AI is applied to actual business operations, enterprises now find themselves needing to consider not only model performance but also how to control and observe AI calls. Building on our experience establishing AI operations environments, CloudNetworks will supply Kong AI Gateway to support the AI governance frameworks of Korean enterprises.
David Carless, Kong's Vice President for Asia Pacific, India, and Japan, said, "This year marks an important juncture where adoption of Kong AI Gateway is gaining momentum among major domestic enterprises in finance, manufacturing, retail, and other sectors. We will further strengthen our cooperation so that companies can build stable AI governance frameworks, leveraging CloudNetworks' mature AI technical capabilities and broad business foundation, and Kong will continue to provide ongoing support for this."
