OpenAI Agents Builder: Build AI Agents (no-code)

Master agent builder basics by setting up your account, understanding pricing, and navigating the canvas, workflows, drafts, and templates in the API platform.

Create an OpenAI platform account to start using the agent builder. Top up credits, verify your organization, and explore templates to build workflows and obtain api keys.

Discover how to access free OpenAI tokens by enabling data sharing for your projects in Agent Builder. Learn to manage per-project sharing, monitor limits, and use these credits for testing.

Explore canvas in OpenAI Agents Builder, learning to create and edit workflows with pen and select tools, test via preview, publish, manage secrets, and duplicate or delete workflows.

Explore core nodes in the agents builder: configure an agent with instructions, prompts, tools (MCP server, file search, web search, code interpreter), and output formats, plus classify and end notes.

Build your first ai agent in a no-code project with agent builder, add instructions and web search, and test it with a practical internet research task.

Explore tools nodes in agent builder, including file search with vector stores, chunking and token concepts. Guardrails block PII, moderation, jailbreak, and hallucination, plus MCP connections.

Master logic nodes, including if/else and branching, using common expression language. Implement while loops and a user approval node to enable human-in-the-loop, multi-agent workflows.

Build a simple personal agent that handles basic queries with a start, agent, and end node, using state variables and a JSON output. Demonstrate moderation with a guard node.

Build a classifier agent with three sub agents to auto classify SaaS customer messages into feature requests, bug reports, or user feedback, using a multi-agent workflow and JSON output schema.

Explore data nodes in the OpenAI agents builder, focusing on transform and set state nodes to manipulate data, manage global workflow variables, and track conversation context using common expression language.

Discover how retrieval augmented generation powers agents with artificial intelligence by fetching context from vector stores and using it to generate accurate answers with private documents.

Learn how vector stores convert text into numeric embeddings, enabling meaning-based search and retrieval of the most relevant chunks for LLM context, using databases like Pinecone.

Two methods create vector stores in the agents builder: upload files via tools file search and name the store, or create outside via storage > vector stores.

Build a customer support agent in the no-code OpenAI agents builder, leveraging a knowledge base and vector store, with a file search tool that automatically handles embedding and retrieval.

Explore widgets in OpenAI Agents Builder, learn to find, create, and embed pre-built visual UI blocks without coding to enhance agent interfaces.

Learn to integrate widgets inside agent workflows by drafting emails and enabling ecommerce product interactions, using either instruction context or a vector store for images and prices.

this lecture demonstrates integrating calendar and other widgets into an ai agent workflow, detailing upload, schema attachment, prompt considerations, and combining widgets to automate tasks and notifications.

Explore an OpenAI document comparison template that analyzes two uploaded PDFs, highlights differences, and presents results via a configurable widget within a multi-agent system.

Build a two-agent workflow with a personal planner and a booking agent that creates a color coded widget, syncs calendar events, uses traffic data, and emails a summary after approval.

Diagnose common widgets issues in OpenAI Agents Builder, from rendering failures and schema mismatches to image loading problems and silent MCP errors.

Publish your agent and workflow to production, obtain a workflow ID, and access it via API calls. Embed with chat kit or agents SDK and test locally before deployment.

Learn essential tools to build ai agents with chatkit, including installing nodejs and visual studio code, cloning the starter template, and running workflows locally before deployment.

Install nodejs and vscode, download and unzip the chat starter pack, configure env local with your OpenAI API key and workflow id, and run the RAC agent locally for testing.

Deploy your RAC agent locally by installing dependencies, configuring environment variables (open api key and workflow id), and running a local server to test the adaptive ai knowledge base.

Create a GitHub account, fork the official OpenAI ChatGPT starter app, and set up a public repo versioned as v1 to prepare for deploying your AI agent.

deploy your agent to a remote server using vercel by connecting your github repo and importing it into vercel's cloud hosting, enabling a live url and domain.

Import your rack agent's GitHub repo and deploy it to production on Vercel, then set two environment variables—the OpenAI API key and the workflow ID—and deploy.

Embed the RAC agent as a chat widget on any website using an iframe embed code, replacing the Vercel URL, and test integration across platforms like WordPress or Webflow.

Walk through OpenAI's ecommerce customer service template, classifying requests into returns, cancellations, or information, and routing them via JSON with conditional paths and human-in-the-loop approvals.

Build a no-code youtube assistant agent that combines a complex prompt, vector store, and web search to generate optimized video content—title, description, pinned comment, tags, and thumbnail concepts.

Build an ultimate email categorizer using MCP with Gmail via Zapier to label emails into categories such as notifications, courses, learning tools, newsletters, and spam.

Implement a classifier-driven, multi-agent workflow that routes inquiries to support, sales, or spam agents for an AI automation agency using a knowledge base and vector store.

Build a multi-agent workflow for a plumbing company that classifies user intent into FAQ or booking, then uses a knowledge base or Google Calendar to answer questions or schedule appointments.

Build a multi-agent system for deep research, combining a keywords agent, a deep research agent, and a report agent, connected via an MCP server and Gmail tool with human-in-the-loop approval.

Customize your AI agent’s chat interface in Agents Builder by adjusting color schemes, accents, typography, and prompts in Widgets Studio playground, then publish as a branded widget.

Discover how the OpenAI Agents Builder's AI generator creates ready-to-use JSON schemas and prompts, automatically typing properties, enums, and arrays for product reviews and task manager apps, saving you time.

Learn to evaluate your no-code AI agent workflows using the evaluate tab, run traces, apply graders, and review results to ensure professional, well-structured responses.

Choose OpenAI models for agents builder by matching use cases like flagship reasoning and long context, then test options from GPT 5.1 to nano for best performance.

Explore observability in agent builder by inspecting logs and traces to see the complete picture of an agent's thought process, including tool calls, data, and outputs in json or text.

Connect agents built in agent builder to Zapier via MCP to access over 8000 apps, configure tools, and automate tasks across Gmail, Slack, calendar, sheets, and more.

Improve cost efficiency for large workflows and automations by applying token strategy, using smaller models, concise prompts, and caching in OpenAI agents.

Finish strong by applying the Agents builder no-code tool and advancing your career in automation and AI agents. Leave a review to help update the course and reach more learners.