Home / Current Issue / Paper 1717873
Speech to Text Generator
Subject area: Science,Engineering and Technology · Area of research: Information Technology
DOI: https://doi.org/10.64388/IREV9I11-1717873
Abstract
In modern educational environments, students often face difficulty in maintaining accurate and complete notes while simultaneously focusing on lecture comprehension. This challenge becomes more significant in multilingual classrooms, where teachers may switch between languages such as English and Hindi during instruction. Manual note-taking in such scenarios is often incomplete, inconsistent, and cognitively demanding. To address this problem, this paper presents a real-time speech-to- text note generation system designed specifically for classroom environments. The proposed system captures a teacher’s speech during live sessions, transcribes it into text in real time, and synchronizes the generated notes across connected student devices. The solution is built using a Flutter-based frontend for cross-platform accessibility, a NestJS backend for session management and synchronization, and the Deepgram API for low-latency multilingual speech recognition. The system supports bilingual speech input, automatic language adaptation, and collaborative note sharing among multiple users within a classroom session. Experimental observations from prototype testing indicate that the system achieves an average transcription accuracy of approximately 80–90% under moderate classroom conditions, with a response latency of 2–5 seconds depending on network quality and background noise. The system reduces the burden of manual note-taking, improves accessibility for students with hearing difficulties, and contributes to more inclusive and efficient learning environments. The work demonstrates the practical application of natural language processing and real-time communication technologies in educational technology solutions.
Keywords
Speech-to-Text, Natural Language Processing, Real-Time Systems, Multilingual Recognition, Deepgram API, Classroom Automation, Educational Technology
How to cite this paper
@article{1717873,
author = {Manas Gajare, Manish Waykole , Purva Mahamunkar, Nishant Patil, Dr. Sanjay Sange; Aruna Kadam},
title = {Speech to Text Generator},
journal = {Iconic Research And Engineering Journals},
year = {2026},
volume = {9},
number = {11},
pages = {1688-1694},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1717873.pdf},
abstract = {In modern educational environments, students often face difficulty in maintaining accurate and complete notes while simultaneously focusing on lecture comprehension. This challenge becomes more significant in multilingual classrooms, where teachers may switch between languages such as English and Hindi during instruction. Manual note-taking in such scenarios is often incomplete, inconsistent, and cognitively demanding. To address this problem, this paper presents a real-time speech-to- text note generation system designed specifically for classroom environments. The proposed system captures a teacher’s speech during live sessions, transcribes it into text in real time, and synchronizes the generated notes across connected student devices. The solution is built using a Flutter-based frontend for cross-platform accessibility, a NestJS backend for session management and synchronization, and the Deepgram API for low-latency multilingual speech recognition. The system supports bilingual speech input, automatic language adaptation, and collaborative note sharing among multiple users within a classroom session. Experimental observations from prototype testing indicate that the system achieves an average transcription accuracy of approximately 80–90% under moderate classroom conditions, with a response latency of 2–5 seconds depending on network quality and background noise. The system reduces the burden of manual note-taking, improves accessibility for students with hearing difficulties, and contributes to more inclusive and efficient learning environments. The work demonstrates the practical application of natural language processing and real-time communication technologies in educational technology solutions.},
keywords = {Speech-to-Text, Natural Language Processing, Real-Time Systems, Multilingual Recognition, Deepgram API, Classroom Automation, Educational Technology},
month = {May},
doi = {https://doi.org/10.64388/IREV9I11-1717873}
}