Function calling
by OpenAI
OpenAI's official playbook for letting models call your tools.
Overview
OpenAI's 'Function calling' documentation explains how to give models access to new functionality and data so they can follow instructions and respond to prompts by interfacing with external systems. It defines the core concepts (tools, tool calls, and tool outputs) and walks through the five-step tool-calling workflow: defining functions with JSON schemas, letting the model decide which to call, executing the call in your code, returning results, and continuing the conversation. The guide also covers namespace organization for grouping related tools, tool search for large tool ecosystems, handling function calls and results, configuration options (tool choice, parallel tool calls, strict mode), streaming tool calls, and custom tools defined with context-free grammars.
At a Glance
- Topic
- Skills
- Level
- Intermediate
- Format
- Documentation
- Cost
- Free
- Duration
- ~25 min read
- Provider
- OpenAI
- Hands-on
- Yes — code/exercises
- Certificate
- None
What You’ll Learn
- ✓How to define functions as JSON schemas the model can call
- ✓The end-to-end tool-calling workflow from model call to returned results
- ✓How to control tool selection with tool choice, parallel calls, and strict mode
- ✓How to organize many tools with namespaces and tool search
- ✓How to stream tool calls and use custom (grammar-constrained) tools
Highlights
- •The canonical, maintained OpenAI reference on function calling
- •Covers strict mode for guaranteed schema adherence
- •Includes scaling patterns (namespaces, tool search) for large tool sets
Who It’s For
Best For
- ✓Developers adding tool use to OpenAI-based apps
- ✓Engineers designing an agent's action layer
Prerequisites
- •Familiarity with the OpenAI Chat Completions or Responses API
- •Ability to read/write JSON schemas
FAQ
What is Function calling?
OpenAI's official Function calling guide, showing how to connect models to external systems and data via JSON-schema tool definitions and a multi-step tool-calling workflow. For developers building tool-using apps and agents on the OpenAI API.
Is Function calling free?
Function calling is free to access.
What level is Function calling for?
Function calling is aimed at a intermediate audience. Recommended background: Familiarity with the OpenAI Chat Completions or Responses API, Ability to read/write JSON schemas.
How long does Function calling take?
Expect roughly ~25 min read. Most learners work through it at their own pace.
What will I learn from Function calling?
You'll learn: How to define functions as JSON schemas the model can call; The end-to-end tool-calling workflow from model call to returned results; How to control tool selection with tool choice, parallel calls, and strict mode; How to organize many tools with namespaces and tool search; How to stream tool calls and use custom (grammar-constrained) tools.