This study aimed to explore the effectiveness of applying the speech-to-text recognition (STR) technology during lectures in English on learning performance of non-native English students. Furthermore, how useful texts generated by the STR for students of different English as foreign language (EFL) ability during different difficulty level lectures was explored. For this reason, one experiment was carried out in this study. Two lectures, one is intermediate difficulty level lecture and the other is advanced, were administered in a traditional classroom and the STR was adopted to aid students' learning. Results of this study showed that students who used STR-generated texts outperformed students who did not use them. According to students, STR-texts were useful to follow the instructor, clarify some words, confirm what the lecturer has said, make up missed information, and enhance understanding. It was found that STR-texts were beneficial for learning of low EFL ability students during both lectures while high EFL ability students only used STR-texts during the advanced level lecture. Based on these results, several suggestions and implications for teaching and research community are proposed in this study.