Comment by sciolist

There's two main ways to "add documents to LLMs" - using documents in retrieval augmented generation (RAG) and training/finetuning models. I believe you can use RAG with Ollama, however Ollama doesn't do the training of models.