Glossary
What is tool calling in AI?
Tool calling, or function calling, is when a language model outputs a structured request to run a defined function, such as search the web, query a database or fill a form, instead of only text. The application runs the function and returns the result to the model. It's how AI agents take actions.
How it works
- The app describes available tools: name, purpose and input schema.
- The model decides a tool would help and returns a call with arguments.
- The app executes the tool and sends back the result.
- The model continues, calling more tools or answering.
Why it matters
Tool calling turns a model from a text generator into something that can act. An AI agent is essentially a loop of reasoning and tool calls until a task is done. Protocols like MCP standardise how tools are offered to many apps.
Good tool design
- Clear names and descriptions; models pick tools from them.
- Simple, well-typed inputs.
- Useful error messages so the model can recover.
- Permissions and confirmations for risky tools.
Example
Asked to check whether a directory listing is live, an agent calls a fetch tool for the listing URL, reads the result, calls a tracker tool to update the status, and reports back.
FAQ
Is tool calling the same as function calling?
Yes, different names for the same capability.
Can models call tools by themselves?
They request calls; the app executes them.
What's a tool schema?
A definition of the tool's inputs, usually JSON Schema.