Last released Aug 30, 2026
RAG evaluation toolkit with readable LLM-judge prompts and local-first (Ollama) support.