Studies report that speech-to-text applications (STT) may support students with writing difficulties in text production. However, existing research is sparse, shows mixed results, and lacks information on STT interventions and their applicability in schools. Therefore, this study aimed to investigate whether a systematic and intensive assistive technology intervention focusing on STT can improve text production. A modified multiple-baseline across-subject design was used involving eight middle school students, four Norwegian and four Swedish. Their STT-produced narrative texts were collected during and after the intervention and the productivity, accuracy, and text quality were analysed. Keyboarding was the baseline control condition. The results demonstrated that seven of the eight students increased text productivity and that the proportion of word-level accuracy was maintained or improved. The use of punctuation progressed in participants with poor baseline skills. Most students’ STT-produced texts had at least a similar ratio of meaningfulness and text quality as keyboarding. However, the magnitude of the changes and development patterns varied, with three students showing the most notable impacts. In conclusion, this study’s intervention seemed beneficial in initially instructing STT, and the progress monitoring guided individually adapted future interventions such as balancing productivity and formal language aspects. Removing the spelling barrier with STT provided an opportunity for students to improve their higher-order skills, such as vocabulary diversity and overall text quality. Furthermore, visible progress, such as the ability to produce longer texts, might motivate continued STT usage. However, such development may not always be immediate.