π-Stack-Web: A Curated Database and Similarity-Guided Platform for Supramolecular Stack Generation

Abstract The generation of stable one-dimensional (1D) supramolecular assemblies is computationally demanding because identifying low-energy stacking configurations requires exploration of complex configurational spaces and highly multimodal energy landscapes. Moreover, even when structurally similar molecular systems have previously been investigated, their optimized stacking geometries are rarely reused systematically, often leading to redundant optimization of closely related supramolecular assemblies. In this work, we present π-stack-web, an online platform for the generation, validation, analysis, and visualization of 1D π-stacked supramolecular assemblies. The foundation of π-stack-web is a carefully curated database comprising 3547 structurally validated and optimized perylene diimide (PDI) and naphthalene diimide (NDI) supramolecular stacking geometries, together with associated stacking parameters and interaction energies. The database spans diverse molecular sizes, side-chain architectures, stacking arrangements, and intermolecular interaction strengths. The platform enables users to construct supramolecular stacks through exact database retrieval or similarity-guided parameter transfer from structurally related molecular systems. The generated assemblies are subsequently evaluated using an automated validation workflow based on steric-clash detection and π-π overlap analysis to assess their physical plausibility. For systems lacking suitable analogues or producing unfavorable stacking arrangements, π-stack-web provides a customized π-stack optimizer package for further offline refinement. By combining a curated supramolecular database with retrieval-guided stack generation and automated structural validation, π-stack-web provides a practical resource for the rapid construction and analysis of supramolecular stacking geometries. The π-stack-web platform is freely available at https://pi-stack-web-app.streamlit.app/