Router Review We built a graph that uses messages as state and a chat model with bound tools. We saw that the graph can: * Return a tool call * Return a natural language response Goals We can think of this as a router, where the chat model routes between a direct response or a tool call based upon the user input. This is a simple example of an agent, where the LLM is directing the control flow either by calling a tool or just responding directly. Let's extend our graph to work with either output! For this, we can use two ideas: (1) Add a node that will call our tool. (2) Add a conditional edge that will look at the chat model output, and route to our tool calling node or simply end if no tool call is performed.
Autonomy
Semi-autonomous
Sandbox-aware
No declared sandbox guidance
Network access
Unspecified
Filesystem access
Unspecified
Permissions declared
Not declared
Pattern
Single agent
Models
gpt-4oclaude-3-5-sonnetgpt-3.5-turbo