Function Calling

Overview

Function calling (also called tool use) lets LLMs invoke external functions to take actions or retrieve information. The model decides when and which tools to call; your code executes them and returns results. This enables true agentic workflows where the model can query databases, call APIs, perform calculations, and more.

Core Concept

The flow is always:

1. Define tools with schemas
2. Send messages + tool definitions to the model
3. Model returns a tool call (name + arguments)
4. Your code executes the function
5. Return the result to the model
6. Model uses the result to continue reasoning
7. Repeat until the model returns a final text response

OpenAI Function Calling