Universal LLM API Proxy & Resilience Library
A self-hosted proxy that exposes OpenAI- and Anthropic-compatible endpoints backed by multiple LLM providers, with automatic API key rotation, failover and rate-limit handling.
Universal LLM API Proxy & Resilience Library is a self-hosted FastAPI service that provides OpenAI-compatible chat-completions and Anthropic-compatible messages endpoints backed by any supported LLM provider, so tools that only support one of those APIs, such as Claude Code or Opencode, can be pointed at other providers like Gemini or OpenAI without code changes.
It bundles two components: the API proxy itself, and a reusable Python resilience library that handles intelligent API key management, rotation and failover across multiple keys and providers, including automatic cooldowns after rate limits or errors. Configured providers can include Gemini, OpenAI, Anthropic and any provider supported by LiteLLM, accessed behind a single API key.
It ships as a downloadable release with an interactive TUI launcher for Windows, a binary for macOS and Linux, and a Docker image, and supports OAuth-based provider credentials stored in a mounted directory.