loading
loading
The ability of an LLM to call external functions and use their results.
When a model supports tool use, you provide a list of available functions with their schemas, and the model can decide to invoke one mid-response. The host application executes the real function and feeds the result back to the model to continue reasoning. This is how agents interact with the real world — reading files, searching the web, querying databases.
Plainly
Think of Tool Use as a named building block in a big LEGO app. The ability of an LLM to call external functions and use their results.
In practice
Use it when you are mapping how the app is structured or explaining a feature to a teammate or agent. In practice, define the owner, input, output, and failure mode before you rely on it.