This Course In a Nutshell

AI has completely changed the role of software developers in the last couple of years, to the point that most commercial software can nowadays be vide-coded.

Our role as developers is no longer to actually write the code ourselves. We have now all been moved by AI into a software architect role, where our main role is now to design the system.

Architectural design decisions, choosing library dependencies, splitting up things into small enough tasks that the AI can do in a way that's easy to verify by a human - that's our role now.

The AI does make us so much more productive - but only those of us who know how to make the most out of it.

Regarding tooling, AI-assisted IDE auto-completion has become almost irrelevant. It's all AI coding agents like Claude Code now, that can simply one-shot medium to large-sized tasks in one go.

So how does this all fit with Angular? With the move to signals and standalone components, and RxJs being optional now, the AI can sometimes have several choices on how to generate code.

The goal of this course is to show you how to correctly configure your favorite coding agent to generate modern, clean, signal-based, and zoneless Angular code.

For that, we are going to be taking Claude Code, and we are going to configure it from scratch to our needs. We will configure the AI with coding recommendations for Angular, Node, Typescript, and we are also going to set up the Angular MCP Server.

The MCP server will allow the AI to get Angular up-to-date documentation, examples, and best practices that match exactly the version of Angular in your project.

Once the initial set-up is in place, we are going to use Claude Code to vibe code an Angular OpenAI chatbot from scratch, using Node and the OpenAI API.

This includes the full generation of all the HTML and CSS in the frontend using Figma screenshots.

The AI chat will be able to answer any Angular question that you have, but it will have guardrails that will prevent its abuse to do other tasks.

You will get all the prompts used, so that you can easily reuse them in your own projects.

We are going to show you how to make the most out of the AI - what level of granularity to use in your prompts, how to verify the AI output, how to prevent the AI from repeating the same mistakes, etc.

Besides the AI chat bot itself, we are also going to show you how to vide-code password based authentication from scratch.

Course Overview

This course is structured around vibe coding one complete full-stack Angular / Node application from scratch.

The application is a signals-based, zoneless Angular AI chat interface backed by a Node and Express REST API that integrates with the OpenAI API. We build both the frontend and the backend together, and nothing is skipped or simplified away.

We begin by setting up Claude Code and learning how to work with it properly: how to provide the right context, how to review what it generates, how to improve its output, and when to challenge or redirect its suggestions.

Next, we move into the heart of the application: the chat interface. We build a collapsible sidebar with conversation history, a polished welcome screen, a live conversation view with typing indicators, and the complete message flow connected to the real OpenAI API through our backend.

From there, we build the full authentication system: sign-up, sign-in, JWT-based session management, route protection, and secure password storage using hashed and salted credentials.

We connect frontend and backend step by step, covering API design, authentication middleware, in-memory data persistence, prompt management, and how to keep system prompts private, secure, and provider-agnostic.

Throughout the course, we apply modern Angular best practices: standalone components, signals, Signal Forms, zoneless architecture, lazy loading, OnPush change detection, and clean separation of concerns between services, components, and models.

Table of Contents

This course covers the following topics:

• Setting up Claude Code and learning how to use it effectively

• Structuring an Angular + Node project for AI-assisted development

• Building a zoneless, signals-based Angular application from scratch

• Backend development with Node, Express, and TypeScript

• Integrating the OpenAI API using plain HTTP requests

• Managing system prompts securely on the server

• Building the chat UI: sidebar, conversation history, and welcome screen

• Creating a collapsible sidebar with active conversation tracking

• The Start / Continue dual endpoint pattern for AI chat completions

• Displaying AI responses with typing indicators

• Global vs local loading indicators and when to use each

• In-memory data persistence with realistic mock seed data

• Authentication middleware and protected routes

• Structured backend logging with Pino

• Environment variable management with dotenv

• Authentication: sign-up, sign-in, JWT sessions, hashed passwords

• Provider-agnostic AI service design

• Using Angular Signal Forms for authentication screens

• Organizing models, services, and components cleanly

What Will You Learn In This Course?

By the end of this course, you will know how to build a complete full-stack AI chatbot application using Angular and Node, and you will have done it while working side by side with an AI coding assistant—so you fully understand how that workflow works in real professional projects.

You will learn how to integrate the OpenAI API securely on the backend, how to design a frontend experience that feels fast and polished, and how to apply modern Angular architecture patterns correctly in a real application.

You will also develop a clear understanding of how to use Claude Code as a professional engineering tool: what to delegate, what to review carefully, how to guide it toward the architecture you want, and how to catch mistakes before they become technical debt.

This is a practical, code-first course built around real applications, real engineering decisions, and real tradeoffs—not slides, toy examples, or theory-only lessons.