Supports efficient RAG research with a Python toolkit that reproduces more than twenty retrieval-augmented algorithms and bundles standard benchmark datasets for evaluation.