The Department of Contemporary History at the University of Vienna hosts a large collection of private papers, correspondences and newspaper extracts mainly from 20th century Austria. An ongoing project aims for the adequate description, storage, processing and publication of these diverse materials (in accordance with their differing legal status). This intention is reflected in a workflow which starts at automated text recognition with Transkribus, goes through exploration experiments in natural language processing and machine-learning, and ends with the issue of sustainable storage (PHAIDRA) and web interfaces apt for the purpose. The talk will present a first sketch of this workflow which is supposed to serve as orientation for future similar endeavours.