Citrix adds NetScaler MCP Gateway controls for LLM and AI agent traffic

Citrix announced new NetScaler MCP Gateway capabilities for governing traffic tied to large language models and agentic AI systems.

The company said the functionality is intended to help enterprises securely route, observe, and control AI traffic between clients, LLM services, and backend Model Context Protocol servers. Citrix’s announcement describes the gateway as a way to apply unified governance to both LLM traffic and agent-to-tool interactions, including model routing and token-level usage tracking.

What Citrix says it is adding

In its July 2026 announcement, Citrix said NetScaler MCP Gateway functionality can securely route and govern agent traffic to backend MCP servers. The company also said the capability supports observability for AI traffic and can help organizations track LLM usage at the token level.

The announcement positions the feature as part of NetScaler’s role in application delivery and security, extending those controls to AI systems that rely on LLMs and autonomous or semi-autonomous software agents.

The Model Context Protocol, commonly shortened to MCP, is used to connect AI applications with external tools and data sources. Citrix’s stated focus is on the network and governance layer around that traffic: deciding where requests go, enforcing policies, authenticating access, and monitoring usage.

Documentation describes a centralized control layer

NetScaler documentation for version 14.1 describes MCP Gateway as a centralized control layer between MCP clients and backend MCP servers. According to the NetScaler Docs page, the gateway supports traffic management, authentication, policy enforcement, and rate limiting.

A related NetScaler deployment guide says NetScaler can operate as an MCP Gateway in either forward-proxy or reverse-proxy mode. The same documentation says requests can be dynamically routed to backend MCP servers using content switching and load balancing.

Those details matter because AI agent deployments often involve many connections between applications, models, tools, and data systems. A centralized gateway gives administrators a place to apply access and traffic rules rather than managing each connection independently.

Why enterprises may care

Citrix is framing the feature around enterprise governance of AI traffic. In practice, that could include routing requests to approved models, limiting how often specific services are called, observing which backend MCP servers are being accessed, and applying authentication before an AI client can reach a tool or service.

The company’s announcement also highlights token-level usage tracking for LLM traffic. Token usage is a key cost and operational metric for LLM deployments, because many model providers price services according to the amount of text processed or generated. Citrix says NetScaler can help track that usage as part of its broader AI traffic controls.

The Silicon Review also reported that Citrix introduced NetScaler MCP Gateway capabilities for unified governance of LLM and agentic AI traffic in enterprise environments, corroborating the broad scope of the announcement.

What is clear from the sources

The sourced materials establish that Citrix has announced NetScaler MCP Gateway capabilities, and that NetScaler documentation describes MCP Gateway as a centralized layer for routing, authentication, policy enforcement, rate limiting, content switching, and load balancing.

The sources do not provide independent customer deployment metrics, pricing details, or performance benchmarks for these capabilities. They also do not establish how broadly the feature is deployed in production environments. For now, the announcement is best understood as Citrix extending NetScaler governance and traffic-management concepts to emerging LLM and MCP-based agent workflows.