ViGPT Researcher: An AI-Powered Online Research Assistant for Comprehensive and In-Depth Analysis

The rapid advancement of artificial intelligence (AI) has significantly transformed the landscape of online research. Large language models (LLMs) have shown great potential in assisting researchers; however, they also present challenges such as hallucinations, bias, and limited contextual depth. This paper introduces ViGPT Researcher, an AI-powered research assistant designed to enhance online research by integrating Retrieval Augmented Generation (RAG) and Plan-and-Solve Prompting (PS Prompting). By leveraging these two techniques, ViGPT Researcher improves information retrieval, minimizes hallucination, and provides users with comprehensive, well-structured research reports. The system is implemented as a web-based application, offering an intuitive user interface and supporting real-time research workflows. This study evaluates the effectiveness of ViGPT Researcher in comparison with existing AI research assistants such as ChatGPT and Gemini. The findings suggest that our approach significantly enhances accuracy, relevance, and user experience in AI-assisted research.