This repository explores the AI Gateway pattern through experimental labs, leveraging Azure API Management to handle AI services APIs with security, reliability, performance, operational efficiency, and cost controls. Primarily focused on Azure OpenAI, the labs offer step-by-step instructions with Jupyter notebooks, Python scripts, Bicep files, and Azure API Management policies to accelerate the experimentation of advanced AI use cases and facilitate the deployment of Intelligent Apps into production.
Key Features
01Model Context Protocol (MCP) experiments
02Backend pool load balancing with Bicep and Terraform
03FinOps Framework integration for AI budget management
04Token rate limiting and metrics emitting
05Agentic experiments with OpenAI Agents SDK and AI Agent Service
06520 GitHub stars
