The biodegradation of synthetic polymers offers a promising solution for sustainable plastic recycling. Polyurethanes (PUR) stand out among these polymers due to their susceptibility to enzymatic hydrolysis. However, the intricate 3D structures formed by PUR chains present challenges for biodegradation studies, both computational and experimental. To facilitate in silico research, we introduce PUR-GEN, a web server tailored for the automated generation of PUR fragment libraries. PUR-GEN allows users to input isocyanate and alcohol structural units, facilitating the creation of combinatorial oligomer libraries enriched with conformers and compound property tables. PUR-GEN can serve as a valuable tool for designing PUR fragments to mimic PUR structure interactions with proteins, as well as characterising simplistic PUR models. To illustrate an application of the web server, we present a case study on selected four cutinases and three urethanases with experimentally confirmed PUR-degrading activity or ability to hydrolyse carbamates. The use of PUR-GEN in molecular docking of 414 generated oligomers provides an example of the pipeline for initiation of the PUR degrading enzymes discovery.