Paper Contents
Abstract
With growing demand for natural and accessible human-computer interaction, voice-based AI systems have become essential in domains such as education, healthcare, and customer service. This paper presents the design and implementation of a real-time AI voice chat application that enables natural, context-aware spoken dialogue. The system integrates Whisper for speech recognition, GPT-based models for natural language understanding, and Tacotron 2 for speech synthesis. Key challenges addressed include low-latency response, multilingual and accent variation, and user data privacy. Our modular architecture ensures cross-platform scalability. Experimental results show a word error rate below 8% and sub-second response times in typical conditions. Limitations such as model drift and speech monotony are discussed, along with strategies for optimization. This work demonstrates a practical, extensible solution for intelligent voice-based interaction
Copyright
Copyright © 2025 Pranjal Amol Badgujar. This is an open access article distributed under the Creative Commons Attribution License.
