<![CDATA[This study aims to develop and test the feasibility and effectiveness of KODIMUSA (Multimodal Digital Corpus of Sundanese Idioms and Proverbs) as a web-based learning platform that integrates multimodal content, gamification, and an artificial-intelligence-based tutor to support the preservation and learning of Sundanese idiomatic expressions. The study employed a Research and Development (R&D) approach using the ADDIE model (Analysis, Design, Development, Implementation, Evaluation). The study involved 65 high school students and undergraduates majoring in Sundanese Language and Literature Education. Data were collected through expert validation sheets, user satisfaction questionnaires, and pretest-posttest instruments. Data analysis utilized feasibility percentage calculations and N-gain scores. Expert validation results showed content validity of 91.6% (highly feasible) and media validity of 88.3% (highly feasible). The pretest-posttest analysis yielded an average N-gain score of 0.60 (moderate), indicating significant improvement in students’ understanding of Sundanese proverbs and idioms. Overall user satisfaction reached 89.2%. Gamification elements, particularly the Snake and Ladder (Oray-Orayan) Pro board game and the XP-based leveling system, significantly increased student engagement. The AI tutor Kang Mus proved effective in providing contextual explanations of Sundanese language expressions. KODIMUSA is a valid, feasible, and effective web-based learning platform for Sundanese babasan and paribasa, contributing to the digital preservation of local cultural heritage while modernizing language learning methods. To strengthen the evidentiary basis, effectiveness was tested inferentially. An independent-samples t-test confirmed that the experimental group’s post-test gain exceeded the control group’s, t (63) = 7.00, p < .001, with a large effect size (Cohen’s d = 1.74), and an ANCOVA controlling for pre-test scores corroborated the result. The 30-item achievement test was reliable (Cronbach’s alpha = 0.89), and ethical clearance was obtained before data collection. KODIMUSA therefore offers a methodologically validated, transferable model for the digital revitalization of endangered regional expressions. The independent-samples t-test confirmed a mean gain difference of 19.6 points (t(63)=7.00, p<.001, 95% CI [14.0, 25.2], d=1.74), indicating a large effect size. Keywords: sundanese idioms and proverbs, digital corpus, gamification, artificial intelligence, web-based learning media.]]>
