Agentic AI, LLM engineering, and the models behind modern automation — multi-agent systems, LLM evaluation and comparisons, RAG, fine-tuning, AI infrastructure, security, and production AI engineering.
From the blog
Learn how to manage conversation history across agent boundaries using nest_handoff_history, per-handoff overrides, CONVERSATION HISTORY blocks, and handoff_history_mapper in the OpenAI Agents SDK.
A deep dive into every parameter of the OpenAI Agents SDK Agent class — from instructions and model settings to dynamic instructions, cloning, and advanced configuration patterns.
Understand the internal agent loop that powers the OpenAI Agents SDK. Learn how agents cycle through LLM calls, tool execution, handoffs, and final output generation.
Learn how to use RunContextWrapper to pass shared state between agents and tools in the OpenAI Agents SDK. Covers typed context, dependency injection, and practical patterns.
Configure the OpenAI Agents SDK for development and production. Covers API keys, model defaults, verbose logging, sensitive data protection, and a production readiness checklist.
Master error handling in the OpenAI Agents SDK. Learn about MaxTurnsExceeded, tool errors, model behavior errors, retry policies, and building resilient agent workflows.
Learn how to use the @function_tool decorator to give OpenAI agents the ability to call Python functions. Covers type hints, docstrings, timeouts, and Pydantic validation.
Master the three execution methods in the OpenAI Agents SDK. Learn when to use async run(), synchronous run_sync(), and streaming run_streamed() with practical code examples.
Build real-time agent interfaces with Runner.run_streamed(). Learn about stream events, WebSocket transport, persistent connections, and building streaming chat UIs.
