This paper presents KeyPoint as an accessibility-first web system for teaching touch typing to blind and low-vision users. The central research problem is that existing mainstream typing tutors rely on visual feedback and are inaccessible to screen reader users, while specialized solutions for the blind lack adaptive algorithms and are often tied to proprietary assistive technology stacks. The technical contribution combines a home-row-anchored learning model, a per-key confidence-based adaptive scheduler that adjusts exercise generation to target weak keys, two audio modes (screen-reader-compatible tonal signals and built-in voice narration), and an assistive-technology-compatible interface path with reduced focus travel and explicit ARIA-based status messaging. A formal per-key metric confidence(key) is introduced to quantify the mastery level of each key and drive the adaptive progression. Analysis of the literature has shown that none of the existing specialized typing tutors for blind users have been the subject of a peer-reviewed scientific publication with architectural description and empirical validation, which positions this work as the first study of its kind. The empirical section is based on a real observational beta test conducted from January 5 to March 15, 2026 and includes 74 users segmented into four accessibility profiles, 623 sessions, 3190 exercises, and 202 assessment points. In the paired T0–T2 cohort (n = 66), mean Correct WPM increased from 19.43 to 29.54 (a 52% improvement), Accuracy from 87.22% to 93.43%, median inter-keystroke interval decreased from 1174.5 to 778.9 ms, and Uncorrected error rate from 4.31% to 0.76% (all p " 0.001, Wilcoxon signed-rank test). Additional analysis revealed that the Blind SR-first profile exhibits 3.6 times more frequent focus losses and 3.7 times longer focus recovery times compared to the Standard profile, empirically confirming that the AT-compatible mode is an essential architectural component rather than an optional feature.
Dmitry Evgenevich Vasilyev (Sun,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: