Search across all documentation pages
8 pages in this section.
Understand how function calling works. Learn the multi-message protocol, key concepts, and when to use this model for agent runtime design or debugging.
Learn tool definitions, call-and-respond cycles, and argument validation with 10 Python examples. Understand tool schemas and agent loops.
Learn to write effective tool descriptions and parameters for models. This guide covers naming conventions, description content, parameter documentation, and evaluation
Define tool input parameters using JSON Schema conventions for OpenAI-compatible APIs. Learn to build robust schemas with examples and best practices.
Learn to handle tool errors in agent loops by catching failures, converting them to structured observations, and enforcing retry budgets for reliable AI.
Learn ten best practices for naming, describing, and scoping tools to ensure agents select them correctly and operate safely.