A new version of the popular software PISA for the analysis of macromolecular interfaces and identification of biological assemblies (complexes) from macromolecular crystal structures is presented. The new web server jsPISA has a substantially improved user interface, based on modern JavaScript technologies, and also new elements of analysis: assembly stock and interaction radar. The new elements help interpretation of PISA results in difficult and ambiguous cases, for example, when the oligomeric state depends on protein concentration, or when the biologically relevant interaction is weak and cannot be easily discriminated from superficial crystal contacts. jsPISA is maintained by CCP4 at http://www.ccp4.ac.uk/pisa. There are no login requirements for using the server.