azure-openai-proxy

Why we included this project

Teams that standardize on the OpenAI API format but are forced to route through Azure OpenAI often find themselves rewriting client code. This proxy handles that conversion for you: it accepts the usual chat completion and embedding requests and forwards them to Azure's endpoint with the right key and header format. Projects like chatgpt-web, chatbox, and LangChain have been tested against it, so existing tooling can keep working without changes. The whole thing is written in Go and ships as a small Docker image, which makes it easy to run as a sidecar or standalone service. It is a focused adapter rather than a full gateway, so treat it as a drop-in translation layer for OpenAI-to-Azure calls rather than a platform.