OpenAI Agents & Orchestration
Official OpenAI examples for building agents: agent orchestration, coding agent with GPT-4.1, reasoning + function calls, multi-agent systems with structured outputs. Real code from the OpenAI team.
Agent Orchestration: Routines and Handoffs
Patterns for building multi-agent systems: routines as step sequences, handoffs for transferring control between agents, implementation from scratch and with the Swarm library.
Coding Agent with GPT-4.1
Building a coding agent with the Agents SDK: apply_patch and shell tools, isolated workspace, user feedback loop.
Reasoning Models + Function Calls
Managing function calling with reasoning models o3/o4-mini via the Responses API: reasoning effort, previous_response_id, handling tool outputs.
Multi-Agent Systems with Structured Outputs
Building a 4-agent data analysis system with Structured Outputs and strict: true: triage, preprocessing, analysis, and visualization agents.
Structured Outputs: Basics and Pydantic
Introduction to Structured Outputs: JSON Schema vs Pydantic parse helper, response_format, function definitions with strict: true.