Building Sarrera: Self-Hosted Enterprise AI Inference Gateway with RBAC, Token Quotas & Telemetry

How to deploy a private, local AI gateway for engineering teams using LiteLLM, Caddy, Langfuse, and Open WebUI with multi-tier quotas.