The COVAT Space OL was developed within the CHC model of intelligence in a collaboration between Cebir and Thomas More (PDC). The test elaborates on the COVAT intelligence tests, which are aimed at children in an educational context. Specifically, it builds on 2 COVAT subtests: Folding Boxes (FB) and Rotated Figures (RF). FB measures visualisation: mental transformation of 2D figures into 3D objects. RF measures speeded rotation: fast mental rotation of simple 2D figures. First we analysed existing data of these classical tests and created new items. To enable remote online testing of adults in a selection context we decided to transform FB into a computerised adaptive test. For RF, adaptive testing was not deemed appropriate because of its emphasis on speed and its item dependency. Therefore, 2 parallel versions of RF were created to enable unproctored testing. The new items were tested in the experimental COVAT Space OLX test (N=167). Rasch analysis was applied to the experimental FB data. This allowed us to create the final FB subtest, with enough items (135) of a wide range of difficulties to estimate a wide range of participant abilities. For the experimental RF subtest (consisting of 3 versions), item times and degrees of difficulty were calculated. Based on the results, 3 new parallel versions were created for the final RF subtest. The 2 subtests of the final COVAT Space OL test each measure 1 narrow cognitive ability of visual processing (Gv). Due to the adaptations to the original subtests, quality assessment of adults' Gv is now possible within selection contexts.