RAGLAB is a modular, research-oriented open-source framework focused on Retrieval-Augmented Generation (RAG) algorithms. It provides reproducibility for six existing RAG algorithms and includes a comprehensive evaluation system with ten benchmark datasets, facilitating fair comparisons between different RAG algorithms and enabling efficient development of new algorithms, datasets, and evaluation metrics.