The prediction of the quaternary structure of biomolecular macromolecules is of paramount importance for a fundamental understanding of cellular processes and drug design. In the era of integrative structural biology, one way of increasing the accuracy of modelling methods used to predict the structure of biomolecular complexes is to include as much experimental or predictive information as possible in the process. We have developed for this purpose a versatile information-driven docking approach HADDOCK (https://www.bonvinlab.org/software) available as a web portal from https://wenmr.science.uu.nl. Despite the past and current success of HADDOCK v2 with >28 900 users and >430 000 jobs processed since 2010, the increasing challenges in modelling today’s large, intricate, and more diverse macromolecular systems require the ability to define custom, highly flexible docking workflows, something for which HADDOCK2.x was not designed. Here, we present HADDOCK v3, a complete rewrite and rethinking of HADDOCK. This new HADDOCK version is a modular software composed of several computational modules that can be combined to create custom-made docking and analysis workflows. Its workflow manager and command-line interfaces are entirely written in Python 3, wrapping around its CNS calculation core, but also integrating already third-party software, thus facilitating its usage as a software suite and a library from within other scientific software. We have focused on versatility, flexibility, scalability, and user experience.