AI Infrastructure
Run agentic AI infrastructure on Kubernetes — agent gateways, MCP, and inference
Start Here: AI Infrastructure 101
What Is an Agent Gateway? The Data Plane for Agentic AI
Learn what an agent gateway is: a unified data plane for MCP, agent-to-agent, and LLM traffic, and why agentic systems on Kubernetes need one.
Installing an Agent Gateway on Kubernetes
Install agentgateway on a Kubernetes cluster with Helm and the Gateway API, then route your first request through it. A hands-on, open-source-only walkthrough.
Your First MCP Gateway
Expose a Model Context Protocol (MCP) server through an agent gateway on Kubernetes, then discover and call its tools through a single governed endpoint.
Routing LLM Traffic Across Providers
Route LLM chat completions through an agent gateway: send traffic to a model you run yourself, then load-balance across multiple providers — all without changing your application.
Agent-to-Agent Communication
Route Agent-to-Agent (A2A) traffic through an agent gateway on Kubernetes, so agents discover and call each other through one governed endpoint.
Securing Agent Traffic with Policy and Observability
Put a security policy in front of agent traffic with agentgateway: enforce API-key authentication on the gateway and read the access logs that record every request.