Recent advances in automatic speech recognition (ASR) and generative AI have enabled new forms of oral language practice beyond traditional, test-oriented speaking tasks. This paper reports on the design and classroom use of a gamified ASR-based oral practice platform developed with Google AI Studio for elementary-level ESL learners. The system integrates ASR into a fast-paced pronunciation game inspired by falling-block mechanics. Learners practice words, phrases, and sentences by speaking aloud: accurately pronounced items disappear, while inaccurate attempts cause items to fall and accumulate as “bricks,” ending the game once a preset height is reached.

The platform was rapidly prototyped using Google AI Studio to manage ASR processing, pronunciation tolerance thresholds, and prompt-based feedback, allowing flexible refinement without complex backend development. The web-based game was piloted in an elementary classroom setting.

Questionnaire data and classroom observations indicate generally positive learner responses, including high engagement, increased willingness to repeat pronunciation attempts, and reduced speaking anxiety compared with traditional drill-based activities. Teachers also reported sustained attention and voluntary practice. Although no quantitative pronunciation gains were measured, the findings suggest that gamified ASR environments can serve as effective supplementary tools for young learners and highlight the pedagogical potential of generative AI for CALL development.