Managing software dependencies is increasingly challenging due to the complexity of modern development, often resulting in “dependency hell” with version conflicts, build failures, and runtime errors. To address these issues, we present AutoPyDep, a recommendation system for Python library dependency management. AutoPyDep features dependency analysis, relationship mapping, and predictive modeling for release categories and dates. By transforming release notes from 23 Python libraries into a graph network, we leverage NLP techniques and a community-based deepWalk algorithm to generate embeddings for tasks such as release category prediction and release date forecasting. Key contributions include a voting classifier achieving a robust F1 score of 0.8 and an ARIMA model with a Mean Absolute Error (MAE) of 1.8 months. AutoPyDep enhances dependency management accuracy, offering actionable insights for developers and supporting improved decision-making in software development. A demonstration of our tool is shared