We present PyNeuralFx, an open-source Python toolkit designed for research on neural audio effect modeling. The toolkit provides an intuitive framework and offers a comprehensive suite of features, including standardized implementation of well-established model architectures, loss functions, and easy-to-use visualization tools. As such, it helps promote reproducibility for research on neural audio effect modeling, and enable in-depth performance comparison of different models, offering insight into the behavior and operational characteristics of models through DSP methodology. The toolkit can be found at https://github.com/ytsrt66589/pyneuralfx.