What is Tool Calling?
Tool calling is how an LLM invokes external functions and APIs — the mechanism by which agents actually do things. It's also where the spend lives.
How it works
The model outputs a structured call; the runtime executes the tool and returns the result.
Tools can be read-only (lookups) or side-effecting (purchases, payments).
Why it matters for cost
Every tool call can be a paid API call.
Retry loops are tool calls repeating the same failure.
The control
Gate side-effecting tools with caps, allowlists, and velocity limits.
FAQ
Is tool calling the same as function calling?
Effectively — function calling is the structured-interface form of tool calling.
Where does the spend risk live?
In side-effecting tools — anything that purchases, pays, or provisions.
Related
Stop the next $12,400 night.
One API call (or MCP tool) in front of every agent transaction — APPROVED, BLOCKED, or FLAGGED, deterministic, ~5 ms, fully logged.
See plans — from $99/mo Try a live check