AIClient2API

This project is a self hosted, multi protocol API proxy that gives large language model applications a single unified interface. It simulates the requests of AI client applications so that model access available through tools such as Antigravity, Codex, Grok, Kiro and Claude can be called as a standard API, and it converts between the OpenAI compatible, Claude and Gemini protocols so existing software can use models it was not written for. It is aimed at developers who want to route coding assistants and other LLM tools through one endpoint. Beyond the client based providers, it supports OpenAI, Claude and custom providers, and adds provider pools, smart routing and automatic failover so requests continue when one account or service fails. The design is modular, so new providers can be added. The documentation site covers installation and Docker deployment, API usage, advanced configuration and a development guide for contributors, in English and Simplified Chinese. The software is free and open source under the GPL 3.0 license. There is no hosted service and no published pricing; users run it on their own machine or server and supply their own accounts. Because it works by presenting client application access as an API, this approach may conflict with the terms of service of some providers, and users should review those terms before depending on it.