This project is a modular LangChain-based application that demonstrates how to build, organize, and run LLM-powered systems using modern AI engineering practices. It covers the full pipeline—from LLM and chat models to embeddings, chains, prompts, and Retrieval-Augmented Generation (RAG)—all organized in a clear and scalable folder structure.

The Bad Improved Prompt Model is designed to enhance user inputs for image generation models. By refining vague or unclear prompts, it helps produce more accurate and visually appealing images.