Function Calling
Allowing AI models to invoke external tools and APIs
What is Function Calling?
The ability of an AI model to generate structured output that triggers an external action, fetching data, calling an API, querying a database, executing code. This is what transforms a language model from a text generator into an agent that can actually do things. Function calling is the bridge between 'AI that talks' and 'AI that acts'. Every agentic workflow depends on it.
How it works
Key takeaways
-
LLM outputs structured JSON instead of plain text
-
Bridge between natural language and APIs
-
Deterministic, same function signature each time
Real-world example
A user says 'Book a meeting with Maria tomorrow at 3pm.' The LLM generates a structured JSON: {action: 'create_event', name: 'Maria', time: '2025-01-15T15:00'}. Your code parses the JSON and calls the calendar API directly.
