In recent years, Retrieval-Augmented Generation (RAG) models have gained popularity in building information retrieval and generative AI applications due to their ability to retrieve relevant information from a knowledge base and generate accurate answers. However, implementing an on-premise RAG service requires not only suitable models but also powerful user interfaces and efficient reasoning frameworks. One such solution is to utilize the Docker-based Open WebUI for easier model management and service integration. In this article, we will discuss how to quickly start Open WebUI using Docker, synchronize Ollama's RAG capabilities, and combine Qwen2.5 models for efficient document retrieval and generation systems.