Last released Oct 8, 2026
Retries, rate limiting and circuit breaking for LLM and HTTP API calls (OpenAI, Anthropic, httpx, requests).