AI capabilities

Last published : Nov 02, 2026
NetScaler® Kubernetes Gateway Controller brings NetScaler AI capabilities to Kubernetes. You configure them with Kubernetes resources, without configuring the NetScaler instance by hand.
NetScaler Kubernetes Gateway Controller supports two AI capabilities:
  • AI Gateway manages and governs traffic between applications and large language model (LLM) backends.
  • MCP Gateway secures and controls traffic between AI agents and Model Context Protocol (MCP) servers.
Related information
AI GatewayMCP Gateway

AI Gateway

AI Gateway sits between your applications and LLM backends, such as Azure OpenAI. It is aware of tokens, so it can route requests to the right model, apply token quotas and rate limits, manage prompts, and move traffic to an alternate model.
For more information, see AI Gateway.

MCP Gateway

AI agents use MCP to call tools on MCP servers. These tools give agents access to systems such as code repositories, databases, and internal APIs. MCP Gateway sits between AI agents and these MCP servers. It routes requests to the right MCP server, controls access to MCP servers, limits tool calls, and handles authentication.
For more information, see MCP Gateway.

Observability

AI Gateway and MCP Gateway can both export analytics to Splunk. Both use the controller's analytics configuration (gatewayController.analyticsConfig in the Helm values), with different analytics profiles for each gateway. If one controller serves both gateways, put the profiles for both gateways in the same analytics configuration.
  • AI Gateway exports per-request AI metrics, such as token counts, model identity, client identity, latency, and cost attribution. For more information, see Observability.
  • MCP Gateway exports a record for each MCP transaction, with details such as the target server, method, tool name, latency, and status. For more information, see Configure Splunk observability.