A language model on its own can only produce text. Tool use (also called function calling) is what lets it act: you describe a set of tools to the model, and when a request needs one, the model replies with the tool's name and its arguments. Your code runs the tool and hands the result back so the model can continue. The model asks for the action; it never runs it itself.
The first lesson explains the concept and the flow of messages between your app, the model, and your tools. The second is a hands-on walkthrough of defining a function, adding it to a chat request, and reading the arguments the model sends back. The last is an Anthropic talk that frames an agent as a model using tools in a loop, and shows why clear tools and context matter so much.
Pay attention to how each tool is described: the name, description, and parameters are what the model reads to decide when to use it. Later modules build on this with planning, memory, and MCP.
IBM Technology's introduction to tool calling: how an app sends messages and tool definitions to an LLM, how the model recommends a tool to call, and how frameworks can manage the loop to reduce wrong calls and hallucinations.
Video 5minA hands-on tutorial using OpenAI function calling: define a function the model can call, add it to the chat request, read the call's arguments from the response, and extend the idea to an appointment-booking example.
Video 4minBarry Zhang of Anthropic on the simplest view of an agent — a model using tools in a loop — and why it helps to think from the agent's point of view, with a limited context window, when designing tools and prompts.
Video 15min