Toolkit for offline RAG application, from creation to synthetic dataset generation and evaluation. Features LLM customization, RAGAS framework metrics evaluation, and result visualization. Takes in custom user documents that will generate QA question-contexts-ground truth pairings and evaluates an LLM's performance in context precision, faithfulness, answer relevancy, and context recall.