cursed-me

Zero-config visual debugging and auto-evaluation for LLM agents. Local-first tracing with a beautiful dashboard for OpenAI, LangChain, CrewAI, and more.