The LLM refers to large deep learning models that are pre-trained on big data. This project enables LLM to access new datasets without further training. It provides a variety of modules required for developing LLM applications, including context source connection and inference based on LLM. This simplifies the development of various applications supported by LLMs, such as chatbots, knowledge bases, and information extraction.