Tool Calling
Giving an LLM a defined set of functions it can request to call, so it can act, not just respond with text.
Reviewed by the RadarTrek editorial team · June 2026
Tool calling (also called function calling) lets you describe available tools — a web search, a database query, a calculator — as JSON schemas the model can choose to invoke. The model never executes the tool itself; it requests a call with specific inputs, your code runs it, and you return the result for the model to continue reasoning with. This separation between reasoning and execution is what keeps tool-using systems auditable.
Why it matters
- —Tool calling is the foundation every agent is built on — without it, a model can only produce text.
- —A tool's description is what the model reads to decide when to call it, so writing a clear description matters more than the implementation.
- —Because your code — not the model — executes every tool call, you can validate, sandbox, and audit every action before it happens.
Where to learn this
Tool Use with Claude
Multi-Agent Systems with Claude course
This is the exact lesson that covers this term in depth — with examples, diagrams, and a hands-on exercise.