So far all the chains/agents in LangChain have been stateless, meaning each interaction is independent This is far from ideal for some real world settings (eg chatbots) where you want the chain/agent to remember previous conversations

→ View original post on X — @langchain