Automatic Project Generation With the Help of LLM Agents

The thesis explores the use of LLM agents for the automatic generation of software projects, analyzing their capabilities, advantages, and risks compared to traditional algorithms and simple LLM queries. The author discusses the architecture of agents, their integration with modern technologies (Django, React, OpenAI API), and application scenarios in business, education, and financial technologies. The system of automatic project generation involving LLM agents is considered. The importance of making an informed choice between complex agents and simple solutions depending on project requirements to ensure efficiency, security, and minimize technical debt is emphasized.