OpenAI Agent Builder & ChatKit: Build & Deploy AI Workflows

Name: OpenAI Agent Builder & ChatKit: Build & Deploy AI Workflows
Rating: 4.4 (25 reviews)

Master OpenAI's Agent Builder platform to create, test, and deploy powerful AI agent workflows with ChatKit integration

Created byPaulo Dichone | Software Engineer, AWS Cloud Practitioner & Instructor

Last updated 10/2025

English

English [Auto],Korean [Auto],

What you'll learn

Build complete AI agent workflows using OpenAI's visual Agent Builder platform without writing complex code
Master all core Agent Builder components including agents, tools, logic nodes, transform nodes, and human-in-the-loop approvals
Design complex multi-agent workflows with while loops, conditionals, and state management
Deploy AI agents to live websites using ChatKit integration and create embeddable chat interfaces
Add guardrails and safety measures to AI agents including user approval nodes and content filtering

Course content

6 sections • 23 lectures • 3h 26m total length

Introduction & Course Prerequisites3:20
Explore OpenAI's agent kit with hands-on practice building AI workflows using a drag-and-drop agent builder, while learning fundamental concepts and leveraging the agent SDK.
What You'll Need for this Course0:32
Open an OpenAI paid account to continue this course and avoid tokens and limits. Opt for a paid plan to avoid tokens and limits.
WATCH DEMO - What You'll Build6:41
Watch a demo of building an AI workflow with agent-based keyword research, user approvals, multi-step research loops, and a summarizer that produces an article, then deploy via ChatKit.

AgentKit and Agent Builder Deep Dive5:28
Discover OpenAI Agent Kit's modular workflow design: build with the agent builder, customize with the agent SDK, and deploy via chat kit, all while integrating prompts, guardrails, and observability.
How To Build AI Agents - Main Components2:08
Build agent workflows with the agent builder canvas, invoking OpenAI models, tools, and guardrails while using memory and a knowledge base to guide logic and customization.
AgentKit Built-in Tools1:23
Explore AgentKit built-in tools that connect to web search, file databases, vector databases, image generation, and code interpretation, plus compute and connectors to MCP servers for richer context.
Hands-on - Analysing a Template Workflow - Nodes and Testing Workflows8:43
Learn to build and test a template workflow in the agent builder, using nodes, edges, and if statements, then preview the flow to verify data and tool interactions.
Hands-on Walkthrough of Each Node and How They Work Internally11:41
Hands-on - Building Our Own Agent Workflow - The Weather or News Agent15:49
Learn to build a grounded classifier agent that routes queries to weather or news workflows using JSON outputs, if statements, and tool-enabled web searches.
Hands-on Adding Widgets to Responses - ChatKit and Widget Builder7:07
Discover how to add widgets to ChatKit responses with the widget builder, create weather widgets, and embed interactive user interfaces on your site using drag-and-drop, no-code workflows.
Hands-on - Creating a Widget for the News Agent7:55
Build a news widget with the widget builder for the generic agent, featuring location header, main news, and three sources, then map the output to a JSON schema.
Hands-on - FAQ Agent with Guardrails6:21
Learn to build an FAQ agent with guardrails by uploading documents to a vector store, enabling a file search tool for semantic retrieval, and enforcing strict source-based answers.
Hands-on - Adding a Classifier Agent and User Approval Node19:27
Hands-on - Evaluation6:55

Hands-on - Adding an MCP Server to the Agent Workflow10:19
Create a new workflow to connect to an MCP server (model context protocol) using the context seven tool, grounding the dev oracle agent with up-to-date docs for code questions.
Hands-on Google Drive MCP Server Setup and Testing5:33
Learn to set up and test a Google Drive MCP server, authorize API access, and use Google Drive as a tool for recent documents, search, and fetch.
Hands-on - Data Nodes - The Transform Node Deep dive13:56
Explore data nodes, including transform and set state, to reshape data, enforce schemas, define global variables, and produce structured JSON outputs within AI workflows, with summarizers and UI widgets.
Hands-on - Set State Node6:26
Create a new flow, add a set state node to assign input text to a topic variable, then interpolate state.topic in the storyteller agent to produce a 150-character tweet.
Hands-on - Research Workflow - Capstone Workflow - The While Loop28:32
build and deploy a full research workflow using set state, while loops, and a keyword researcher agent to generate keyword phrases and use the web search tool.

Requirements

An OpenAI account with access to the Agent Builder platform
No coding experience required - this course is designed for visual builders, but basic coding experience will be helpful

Description

OpenAI Agent Builder is here.

Transform your ideas into powerful AI applications without writing complex code!

Welcome to the most comprehensive course on OpenAI's revolutionary Agent Builder platform and ChatKit deployment system. Whether you're a business owner, entrepreneur, developer, or AI enthusiast, this hands-on masterclass will empower you to create sophisticated AI agent workflows that can automate complex tasks, process information, and interact with users through beautiful interfaces.

What makes this course unique?

This isn't just another AI theory course. You'll build real, deployable AI agents from scratch, including:

An intelligent article writer and summarizer with human approval workflows
A weather and news aggregation agent with custom widgets
An FAQ agent with guardrails for safe, controlled responses
A complete research workflow using while loops and multiple specialized agents
Live deployment of your AI agents to websites using ChatKit

By the end of this course, you'll have:

Complete mastery of OpenAI's Agent Builder visual interface
Multiple working AI agent workflows ready for production
Your own deployed AI chatbot on a live website
The skills to create custom AI solutions for any business need
Deep understanding of MCP servers, data transformation, and agent orchestration

With over 22 hands-on lectures covering everything from basic node connections to advanced MCP server integration and widget creation, you'll gain practical experience with every aspect of the platform.
Learn how to add web search capabilities, file processing, code interpretation, and create beautiful custom UI widgets for your AI responses.

Stop struggling with complex coding frameworks.

Start building production-ready AI agents today with the visual power of OpenAI Agent Builder!

Who this course is for:

Business owners and entrepreneurs who want to create custom AI solutions without hiring developers
Product managers and designers looking to prototype AI features quickly
Developers who want to rapidly build AI workflows without extensive coding
Anyone interested in the no-code/low-code AI revolution who wants to build real, deployable solutions
Technical and non-technical professionals who need to automate workflows with AI

OpenAI Agent Builder & ChatKit: Build & Deploy AI Workflows

What you'll learn

Explore related topics

Course content

Introduction3 lectures • 11min

Code and Resources1 lecture • 4min

OpenAI's Agent Kit - Agent Builder and ChatKit Deep Dive11 lectures • 1hr 33min

AgentBuild and MCP Servers and the Transform Node5 lectures • 1hr 5min

Deploying Agents Workflow and ChatKit2 lectures • 32min

Wrap up and Next Steps1 lecture • 2min

Requirements

Description

Who this course is for: