Easy Structured Output from LLMs As more and more models support function calling, this is becoming the default way to get structured outputs from LLMs. We've added a new method on LangChain LLMs to make using this as easy as possible. Examples of model providers supporting