This service uses langchain to vectorize documents into a FAISS vectorstore.
The embdedding model used is BAAI/bge-large-en-v1.5 available in Huggingface.
These vectorstores can then be used to do similarity searches in this vector space. This process is used in RAG (Retrieval Augmented Generation) to augment the generation process of a LLM with information from the document.