Recent advances in Recommender Systems (RSs) have led to the development of numerous software frameworks that facilitate research, evaluation, and deployment. However, these tools vary significantly in their implementations, supported algorithms, evaluation strategies, and usability, making it challenging for researchers to select the most suitable framework for a given task. This paper presents a structured comparative study of RS software frameworks, focusing on their support for recommendation techniques, dataset compatibility, and evaluation metrics. We classify several tools based on different perspectives: the programming languages in which they are implemented, recommendation techniques (e.g., collaborative filtering, content-based, hybrid), dataset compatibility (e.g., MovieLens, Amazon, Netflix), and evaluation metric support (e.g., RMSE, NDCG, Precision). Our analysis serves as a practical guide for researchers and developers in selecting appropriate RS software based on task requirements and evaluation objectives.