Loading content
The model emits a structured call — a named function with typed, schema-validated arguments — that the harness executes, feeding the result back into the model's context; the model decides which function to call, when,… Typical conditions for use: External actions or data sources must be integrated in a controlled manner; Arguments need strict validation. Th…
Use this profile to understand the building block briefly, place it in the model, and switch to the 360° assessment when needed.
Theoretical construct: explains a term, principle, or mental model.
What organizes, connects, or makes decisions possible.
Function Calling is a pattern in which a model does not generate free text for a tool, but emits a named, structured call with typed, schema-checked arguments.
In LLM-based agent systems, Function Calling emerged from the problem of connecting external actions and data sources in a controlled way instead of parsing free text after the fact. The pattern separates model and execution: the model chooses an allowed function and forms valid arguments; a harness checks the schema, runs the call, and feeds the result back into context.
Think of Function Calling as a gate between reasoning and execution. The model decides which action is needed, but it may send only the function name and arguments. The schema defines the allowed shape, the harness validates the payload, launches the tool, and returns the response. That keeps action bounded while still letting the system act.
The model chooses only from approved functions; the name determines which action can run.
Arguments must satisfy the contract's types, required fields, and rules before execution.
The runtime checks, dispatches, and oversees the call outside the model.
A protocol for standardized access to external tools and resources.
Entire agents can also appear behind the same tool interface as functions.
A contrasting approach where executable code, rather than structured function calls, is the primary action medium.
Function Calling is useful when a model should steer a calendar, search, database, or workflow only through clearly bounded actions. It reduces parsing errors and makes calls inspectable, but it adds schema maintenance, integration effort, and referent validation. Type-correct but invented values can still cause harm if the runtime accepts them.
Where this building block is located in the topic model.
No structure path available.
Explore how this building block connects to concepts, methods, technologies, and tools.
These sources establish the term and its professional meaning.
All direct connections of the current building block in a compact text view.
This classification shows where the building block typically matters, how demanding it is, and what kind of impact it has in the model.
The level within the organization (enterprise, domain, team) at which the AssetBlock is applied.