Tool Schema
The spec sheet that tells the model what a tool does, what it’s called, and exactly what inputs it takes.
Think of it like
The label on a control panel — button name, what it does, and which knobs it needs set first.
Example
A `search(query, max_results)` schema tells the model the tool’s name, that `query` is a required string, and `max_results` an optional number.
How it actually works
Each tool is described in a structured format — usually JSON Schema — with a name, a natural-language description, and typed parameters. The model reads these descriptions to decide which tool fits and how to fill its arguments, so the wording matters as much as the types: a vague description means the model misuses the tool. It is the contract between the model’s intent and your code.
For product teams
Half of tool reliability is just writing clear tool descriptions.
For engineers
A machine-readable declaration (typically JSON Schema) of a tool’s name, description, and parameter types that constrains the model’s call.
Related
- Function Calling — What the model produces to satisfy the schema.
- Tool Use — The broader capability the schema enables.
- Structured Tool Use — A stricter form that validates the call against the schema.
Read anything AI without the jargon
Look up any term in plain English, or save terms as you read with the free Chrome extension.
Open DecoderAdd to Chrome