pretty easy to swap this out! we have standard interfaces for most llm/embeddings providers

→ View original post on X — @langchain