The proposed Real-Time Indian Language Translator addresses the socio-linguistic complexities of the Indian subcontinent, where diverse regional dialects often impede efficient inter-state communication. This research introduces a framework that integrates NLP and Deep Learning models to achieve low-latency translation between multiple Indic languages. The system architecture supports multimodal inputs—specifically speech-to-text and text-to-text—utilizing automated language detection and neural processing to generate accurate results. By optimizing translation reliability and speed, the system aims to democratize access to essential services such as digital governance, emergency healthcare, and inclusive education. The primary contribution of this work is the development of a robust, efficient medium to bridge the gap between speakers of disparate linguistic backgrounds within the country.

The Real-Time Indian Language Translator successfully addresses the challenges of linguistic fragmentation by providing a seamless, high-fidelity translation framework for the Indian subcontinent. By synthesizing Automated Speech Recognition (ASR), Neural Machine Translation (NMT), and Text-to-Speech (TTS) into a unified architecture, the system achieves near-instantaneous cross-lingual communication. A key contribution of this research is the system’s dual-mode versatility, which ensures reliable performance in both high-connectivity urban centers and remote, offline environments. The modularity and scalability of the proposed model suggest significant potential for integration into critical public infrastructures, such as e-governance, tele-healthcare, and inclusive pedagogy, ultimately fostering a more linguistically integrated society.

[1] Radford, A., Kim, J. W., Xu, T., Brockman, G., McLeavey, C., & Sutskever, I. (2023). “Robust Speech Recognition via Large-Scale Weak Supervision.” OpenAI Whisper Technical Report. [2] Gala, J., Chitale, P. A., et al. (2023). “IndicTrans2: Towards High-Quality and Accessible Machine Translation Models for all 22 Scheduled Indian Languages.” Transactions on Machine Learning Research (TMLR). [3] Bhogale, K., et al. (2023). “Vistaar: Diverse Benchmarks and Training Sets for Indian Language ASR.” Proceedings of INTERSPEECH 2023. [4] Tripathi, K., et al. (2025). “Enhancing Whisper’s Accuracy and Speed for Indian Languages through Prompt-Tuning and Tokenization.” IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). [5] Soni, A. A. (2025). “Improving Speech Recognition Accuracy Using Custom Language Models with the Vosk Toolkit.” arXiv preprint arXiv:2503.21025. [6] Chandra, R., et al. (2025). “An Evaluation of LLMs and Google Translate for Translation of Selected Indian Languages via Sentiment and Semantic Analyses.” IEEE Access, Vol. 13. [7] Nair, V. (2025). “AI-Powered Language Translation System leveraging Neural Machine Translation.” Journal of Integrated Engineering Sciences, 1(2), 31-39. [8] Dabre, R., et al. (2025). “Towards Building Large Scale Datasets and State-of-the-Art Automatic Speech Translation Systems for 14 Indian Languages.” Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (ACL). [9] Google Cloud. (2024). “Translation API Product Documentation.” Available at: https://cloud.google.com/translate/docs. [10] Vosk Speech Recognition. (2023). “Vosk Toolkit: Offline Speech Recognition for 20+ Languages.” Available at: https://alphacephei.com/vosk/.

© 2025 International Journal of Engineering and Techniques (IJET).

Digital Object Identifier (DOI)

IJET assigns a unique DOI to every accepted article via Zenodo for persistent linking and citation. Your article’s DOI: https://doi.org/{{doi}}

Open DOI About Zenodo DOI

Close