Understanding LLM Routers' Role in Agent Architectures
Aparna D. wrote up a post here on LLM routers in agents. These are hugely important to agent architectures. Feel free to drop a note if you are using LangGraph framework vs just LLM routers in your app? .These routers are made of:
A LLM call using function aka tool calling to choose between a number of different skills to use depending on user input
Skills or functions that complete the work of the agent. E.g. another LLM call, RAG response, or an internal API call
Memory to store variables accessed by both the router and skills
