Створення інтелектуального веб-асистента для аналізу даних з кульової стрільби на основі LLM

Abstract

The development of an intelligent web application for the collection, aggregation, and analysis of up-to-date information on target shooting using large language models (LLMs) is considered. The proposed system combines RetrievalAugmented Generation (RAG) technology, deep web scraping of official sources, and global web search to generate comprehensive reports in real time. The integration of the Gemini 2.5 Flash model via API has been implemented, and an interactive chat module with session context retention has been developed based on the Streamlit framework. The hybrid search architecture, main functional modules, and software testing results are described.