Early detection of reading acquisition failure is crucial since intervention is most effective if started early. This presentation gives a cross-linguistic overview of studies, conducted at our department, making use of GraphoLearn (GL), a serious gaming environment, in three languages with relatively transparent orthographies. Aim of the studies is to evaluate the effectiveness as well as screening value of playing GL. GraphoLearn was used to develop early reading acquisition games in three languages: Spanish, Standard Indonesian (both highly transparent orthographies) and Dutch (intermediate transparency). In all four studies typically developing children, and in some studies, children with developmental language disorder (DLD) and/or reading difficulties participated in a research protocol consisting of a pre-test at the start of reading education, followed by a period of playing GL, and concluded with a post-test and follow-up. The pre- and post-tests comprised standard sets of preliteracy tests. The post-tests also included word decoding tests. Playing consisted of matching (strings of) graphemes on a computer screen with spoken fragments. Results show a mainstream learning trajectory (strong regression) from preliteracy skills, via accuracy and speed of letter-sound association during playing, to early (post-test) and later (follow-up) word and pseudoword decoding. Patterns of qualitative (e.g. errors) and quantitative (e.g. learning rate) in-game data predicting and influencing later reading fluency were found for different subgroups of children. We conclude that GL is an effective tool for early dynamic diagnostic screening, and supporting early reading acquisition. Future perspectives targeting integrated, multi-factorial and cross-modal digital game-based learning (DGBL) are discussed.