Pronunciation is often seen as the most anxiety-inducing aspect of language learning. Various techniques have been introduced to improve students’ pronunciation, including explicit instruction (Zhang & Yuan, 2020), virtual reality (Alemi & Khatoony, 2020), and speech-to-text technology (Jiang et al., 2021). As explicit instruction requires instructor expertise, speech-to-text technology has recently gained increased attention, even though free versions have been perceived as inferior. This quasi-experimental study examined the effectiveness of free speech-to-text software in improving Japanese university students’ spoken English. A pre-test-post-test comparison was conducted with two groups (N = 77) over 15 weeks, with both groups receiving explicit pronunciation instruction via flipped learning videos. During weekly classes, the control group (n = 28) practiced listen-and-repeat exercises, while the treatment group (n = 49) used free speech-to-text software. Based on recorded speeches rated by three evaluators (α = .800), the treatment group showed statistically significant pronunciation improvement (p = .003). While both groups improved in rhythm, intonation, and intelligibility, the treatment group exhibited greater gains, though not at a significant level (p > .05). These findings suggest that flipped learning enhances pronunciation instruction, with additional benefits when combined with speech-to-text technology.