Kick off your journey with the OpenAI Agents SDK by learning how to run your first agent, explore its results, and format outputs effectively. This course also introduces asynchronous and streamed execution to give you flexibility in building responsive applications.
Overview
Syllabus
- Unit 1: Getting Started with OpenAI Agents in JavaScript
- Modify Agent Instructions to Create an Italy-Only Travel Assistant
- Choosing a Specific Model for Your Agent
- Configuring Model Parameters for Concise Agent Responses
- Debug and Fix the Agent Output
- Using Runner Instances for Agent Control
- Create a Joke-Telling Agent
- Unit 2: Inspecting and Understanding Agent Result Properties
- Accessing Result Object Properties
- Exploring Agent Steps with newItems
- Filtering Agent Messages
- Examining Raw Responses from the Language Model
- Tracking Token Usage and API Metrics
- Unit 3: Structuring and Type-Checking Agent Outputs with Zod
- Structured Output with Zod Schema
- Fix Type Mismatches in the Schema
- Extending Your Agent's Output Schema
- Extracting and Printing Structured Agent Output Fields
- Create a Custom Travel Budget Schema
- Unit 4: Streaming Agent Responses
- Enable Streaming Mode for Agents
- Implement Real Time Text Streaming
- Manual Event Processing Implementation
- Custom Text Streaming Implementation
- Using Runner Class for Streaming