Elementary science lessons on classifying animals by food type often rely on static media that offer limited opportunities for interaction. This study aimed to describe the development of augmented reality (AR) learning media built on the Assemblr Edu platform using the ADDIE model, and to examine whether and how the media meets criteria of validity, practicality, and effectiveness among third-grade students. A research and development design was applied with 16 third-grade students, one material expert, one media expert, and one classroom teacher. The instruments were expert validation sheets, teacher and student response questionnaires, and an evaluation test, and the data were analyzed descriptively as percentages. The material expert and the media expert gave scores of 91.11% and 82.5%, respectively, both in the very valid category. The teacher response reached 91.11% (very practical) and the student response reached 85.63% (practical). A total of 13 of 16 students (81.25%) reached the minimum completion criterion, with a class mean of 83.6, which met the effectiveness criterion set in this study. These findings suggest that the media is feasible for classroom use as a supplement to science instruction, although the small sample, the absence of a comparison group and a pretest, and the delivery of the lesson by the researcher limit generalizability. Future studies should apply pretest-posttest designs with comparison groups, teacher-led implementation, and longer instructional periods.
Copyrights © 2026