Structure-based drug design (SBDD) requires the coordinated execution of multiple computational tasks. Although many powerful software tools are available for each individual step, practical SBDD workflows often remain fragmented, requiring substantial manual intervention, technical expertise, and repeated data conversion between heterogeneous programs. This reduces accessibility for non-expertise users. Here, we present an agentic AI framework that automates a series of SBDD tasks using LangGraph. Rather than introducing new softwares, our approach focuses on the intelligent integration and automated execution of established computational tools. The pipeline sequentially performs protein structure search from the Protein Data Bank (PDB), protein structure prediction (co-folding), binding site prediction, molecular docking, rescoring of docking poses, and visualization of protein–ligand interactions. The framework is designed to reduce manual intervention in multi-step computational drug discovery. Since the framework is modular, individual tools can be replaced or extended depending on the research purpose. Our framework suggests that agentic AI can serve as a practical strategy for streamlining complex structure-based drug discovery workflows in real-world research settings.