This project is based on converting an audio character pool to text using the speech-totext API. Speech-to-text conversion includes secondary, secondary, and primary languages. The process is like speaking and transcribing. This article provides a comparison of the strategies used in small, medium and large speakers. Comparative studies have identified the advantages and disadvantages of each method to date. An attempt to demonstrate the power of language structure in improving speech-totext accuracy. We conducted experiments on speech data with loud and incomplete sentences. The results showed that the selected sentences were more effective than the linked sentences.