This presentation introduces an approach towards full-text digitization of Republican China newspapers within the Early Chinese Periodicals Online project (ECPO) . Since full pages cannot yet be automatically processed by OCR engines (dense document layout, special characters, imperfect visuals), we chose two methods that can later be combined. Firstly, we implemented an annotation tool to create, label and group bounding boxes, defining e.g. an article. In an OCR pipeline we process these segments and evaluate different OCR engines. The segments will also be used to train a neuronal network for -future- automatic layout recognition. Secondly, we established a double-blind keying workflow to create high-quality full-text. With the TEI XML texts we build a dictionary to train and improve OCR. Our outcomes are quality ground truth data sets (texts and segmentation: GitHub), an OCR-engine evaluation, and a full text data-set for Republican China newspapers, searchable within ECPO.