Reference

Tech Stacks Explained

You do not need to master every stack before you start building. You need to know what each one does, who uses it in the real world, and when to choose it. AI fills in the rest. This page is that reference.

Jump To

Stacks Covered

Frontend

HTML / CSS / JS  ·  Bootstrap  ·  React  ·  Vite

Backend

Node.js / Express  ·  Python  ·  Ruby on Rails  ·  C#  ·  PHP  ·  Java

Data & APIs

SQL / NoSQL  ·  Firebase  ·  REST APIs  ·  GraphQL

Frontend

The Presentation Layer

What users see and interact with. Frontend runs in the browser. These are the technologies that render your UI.

01

HTML5 / CSS3 / Vanilla JavaScript

What it is: The foundation of every website on the internet. HTML is the structure (the bones). CSS is the style (the skin). JavaScript is the behavior (the muscles). No framework, no build step, no install. Open a file and go.

Good for: Marketing sites, documentation, portfolios, landing pages, informational platforms. Anything that does not need complex state management or a dynamic data layer. This entire site is built on it.

Who uses it: Every company on the internet has some vanilla HTML/CSS somewhere. Cloudflare, Wikipedia, most content sites, most landing pages. When a team wants something that loads fast and is easy to maintain, they go back to basics.

VS Code extensions:

Live Server — instant local dev server with hot reload
Prettier — auto-format HTML, CSS, JS on save
CSS Peek — jump from class name to CSS definition
Auto Rename Tag — rename opening and closing tags together
HTML CSS Support — class and ID completion

Choose it when: You are building a content site, a portfolio, or anything that does not need a database or user accounts. Zero dependencies means zero things to break. Claude Code handles it perfectly with no configuration.

Billy's Take This site is built entirely in HTML/CSS/Vanilla JS. No React. No Node. Fastest load times possible, easiest to deploy, and Claude Code navigates it without getting confused by build tools.
02

Bootstrap

What it is: A CSS framework that gives you a grid system, pre-styled components (buttons, cards, modals, navbars), and responsive utilities out of the box. Link one stylesheet and your site is 80% of the way styled before you write a single custom rule.

Good for: Rapid prototyping, admin dashboards, internal tools, anything where you need it to look professional fast and do not have time to design a custom system. It is not the most distinctive-looking CSS, but it is functional and consistent.

Who uses it: Twitter originally built Bootstrap and open-sourced it. LinkedIn, Spotify's internal tools, countless government sites, startups that need to ship fast. It has the largest adoption of any CSS framework in history.

VS Code extensions:

Bootstrap 5 Quick Snippets — autocomplete for BS5 classes
IntelliSense for CSS class names — completion in HTML files

Choose it when: You need to build something that looks clean in days, not weeks. Use a custom CSS system (like this site) when you want a distinctive visual identity. Use Bootstrap when you need working and consistent fast.

Current status: Bootstrap 5 is the stable version as of 2026. Bootstrap 4 is still deployed everywhere — you will see it in legacy codebases. It is not legacy itself but custom CSS frameworks have taken mindshare at the high end.

03

React

What it is: A JavaScript library from Meta for building component-based user interfaces. Instead of writing HTML directly, you write reusable components — self-contained chunks of UI with their own logic and state. React re-renders only the parts of the page that changed, making complex apps fast.

Good for: Single-page applications, dashboards, social feeds, anything with lots of interactive state. If you are building something where the UI updates constantly based on user actions or live data — React is the industry standard.

Who uses it: Meta (Facebook, Instagram, WhatsApp), Airbnb, Uber, Netflix, Dropbox, Pinterest. The question is easier to answer as who does not use React. It has dominated frontend development since 2016.

VS Code extensions:

ES7+ React/Redux/React-Native snippets — instant component boilerplate
Prettier — auto-format JSX
ESLint — catch React-specific errors before runtime
React Developer Tools — browser extension for inspecting component trees

Choose it when: You are building an app with complex state, lots of interactivity, or a team of developers. React has the largest ecosystem, the most tutorials, and the most job postings of any frontend technology. If you are learning one JavaScript framework, learn React.

Note React has a learning curve that HTML/CSS does not. Do not start with React on your first project. Build something small in vanilla JS first, understand what the problem is, then React will make sense as the solution.
04

Vite

What it is: A build tool and development server for modern JavaScript projects. Where the old standard (Create React App) took 30+ seconds to start and minutes to build, Vite starts in milliseconds. It uses native ES modules during development so it never has to bundle everything upfront.

Good for: React projects, Vue projects, any modern JavaScript app that needs a fast dev environment. Vite is now the default recommendation when starting a new React project — CRA is effectively retired.

Who uses it: Shopify, GitLab, and the broader JavaScript ecosystem. Vite was built by Evan You (creator of Vue.js) and has become the standard bundler for new projects across all major frameworks.

VS Code extensions:

Vite (official) — run Vite commands from the VS Code command palette

Start a React project with Vite:

npm create vite@latest my-app -- --template react
cd my-app
npm install
npm run dev

Choose it when: You are starting a new React or Vue project. There is almost no reason to use Create React App in 2026. Vite is faster to start, faster to build, and better maintained.

Backend

The Logic Layer

Backend runs on a server. It handles business logic, database queries, authentication, and anything that should not happen in a user's browser.

05

Node.js / Express

What it is: Node.js lets you run JavaScript on the server — not just in the browser. Express is the most popular web framework for Node.js. It handles routing, middleware, and HTTP requests with minimal setup. Together they are the most commonly deployed JavaScript backend stack.

Good for: REST APIs, real-time applications (chat, live feeds), microservices, command-line tools, anything that needs to handle many concurrent connections without blocking. The same language on frontend and backend is a big developer experience win.

Who uses it: Netflix (moved to Node to cut startup time by 70%), LinkedIn (rewrote mobile backend in Node, 10x server reduction), PayPal, Walmart, NASA. Vercel's serverless functions run on Node.js by default.

VS Code extensions:

Node.js Extension Pack — debugging, snippets, npm scripts
REST Client — send HTTP requests directly in VS Code
Thunder Client — lightweight API testing like Postman inside VS Code
DotENV — syntax highlighting for .env files

Choose it when: You already know JavaScript and want to use it on the backend. Or when you are building APIs that will be consumed by a React frontend. JavaScript full-stack (Node backend + React frontend) is the most in-demand web development skill combination in 2026.

06

Python

What it is: A general-purpose programming language known for readable syntax and an enormous standard library. Python is the dominant language for data science, machine learning, automation, scripting, and AI/ML research. It is also used for web backends via Django and Flask.

Good for: Data science and analysis (pandas, numpy), machine learning and AI (TensorFlow, PyTorch, scikit-learn), automation scripts, web scraping, APIs, rapid prototyping. If you are building anything that touches AI/ML, Python is not optional — it is the language of that entire ecosystem.

Who uses it: Google (built many internal systems in Python), Instagram (Django backend), Spotify (data pipeline), Dropbox (desktop client), NASA, CERN. OpenAI's API client libraries are Python-first.

VS Code extensions to make it feel like PyCharm:

Python (official Microsoft extension) — interpreter, debugging, IntelliSense
Pylance — fast type checking and code intelligence
Jupyter — notebook support directly in VS Code
Black Formatter — auto-format Python on save (PEP 8 compliant)
Ruff — extremely fast Python linter, replaces flake8 and isort

Choose it when: You are doing anything data-related, building AI/ML pipelines, writing automation scripts, or want to get into the AI/ML field professionally. Python is also great for beginners — the syntax is close to plain English and error messages are readable.

AI/ML Note Every AI framework that matters — TensorFlow, PyTorch, Hugging Face, LangChain, CrewAI — has Python as the primary interface. If you want to build AI-powered applications beyond using an API, Python is where that happens.
07

Ruby on Rails

What it is: A full-stack web framework built on Ruby. Rails follows "convention over configuration" — the framework makes decisions for you so you can build faster. A developer who knows Rails can build a functional web app in hours that would take days in a lower-level framework.

Good for: Rapid prototyping, startups, content management, e-commerce, B2B SaaS. Rails is opinionated in a way that eliminates decision fatigue and lets you focus on what the app actually does.

Who uses it: Shopify (one of the largest Rails deployments in the world), GitHub (originally built on Rails), Basecamp (37signals invented Rails), Airbnb (early version), Twitch (early version), Hulu. Rails powers a substantial portion of the internet's e-commerce and SaaS layer.

VS Code extensions:

Ruby LSP (Shopify official) — language server, IntelliSense, debugging
Rails (by bung) — snippets for controllers, models, views
ERB Formatter/Beautify — format ERB templates

Choose it when: You want to build a full-stack web app with a database, user authentication, and admin features as fast as possible. Rails does not have the raw performance of Go or the ecosystem of Node.js, but it has the best developer productivity per hour of any web framework for building standard web apps.

08

C# (.NET)

What it is: A statically typed, object-oriented language from Microsoft. C# runs on the .NET platform. It is used for enterprise web applications (ASP.NET), Windows desktop applications, and is the primary language for Unity game development.

Good for: Enterprise software, Windows applications, Unity games, large-scale ASP.NET web apps, Azure cloud services. C# has the best tooling in the Microsoft ecosystem and the strongest typing system of the commonly-used languages.

Who uses it: Stack Overflow (runs on .NET), Microsoft (obviously), Xbox (Unity games), most large enterprises with Windows-native software, banking systems, healthcare applications. Corporate environments with Azure infrastructure run heavily on C#.

VS Code extensions:

C# Dev Kit (Microsoft official) — full IntelliSense, debugging, project management
.NET Install Tool — install and manage .NET SDKs
Unity (if doing game dev) — Unity engine integration

Choose it when: You are building for Windows, building with Unity, or working in a corporate environment that runs on Microsoft infrastructure. C# pays well — enterprise companies desperately need people who can navigate legacy .NET codebases.

09

PHP

What it is: A server-side scripting language that has powered the web since the late 1990s. WordPress is written in PHP. So is Laravel, the modern PHP framework that gives PHP a clean, expressive syntax and a modern development experience.

Good for: WordPress sites and plugins, Laravel applications, legacy web system maintenance, shared hosting environments. PHP runs on more web servers than any other backend language — because WordPress runs on PHP and WordPress powers ~43% of all websites.

Who uses it: WordPress.com, Facebook (originally built in PHP, now uses HHVM), Wikipedia, Slack (parts), Etsy (originally PHP). Every hosting provider on earth supports PHP. The barrier to deploy a PHP app is the lowest of any backend language.

VS Code extensions:

PHP Intelephense — best PHP language server (paid license for full features)
PHP Debug — Xdebug integration for step-by-step debugging
Laravel Extra Intellisense — autocomplete for Laravel apps
PHP DocBlocker — auto-generate PHPDoc comments

Choose it when: You are building on WordPress, maintaining a legacy PHP system, or using Laravel for a new project. PHP has a reputation for messy old code — and that reputation was earned. Modern Laravel PHP is a different experience. Knowing PHP opens a huge job market of maintenance and WordPress work.

Legacy Note PHP was voted "most dreaded language" for years on Stack Overflow surveys. Modern Laravel PHP is genuinely good. The dreaded part is the 20-year-old PHP 4/5 code that still runs half the internet. Knowing how to deal with it pays well.
10

Java

What it is: A statically typed, object-oriented language built by Sun Microsystems (now Oracle). Java's "write once, run anywhere" JVM (Java Virtual Machine) means code compiled once runs on any system. It has been the dominant enterprise backend language for 25+ years and powers most Android applications.

Good for: Enterprise software at scale, Android development, financial systems, large-scale microservices (Spring Boot), anything that runs inside a massive corporate infrastructure. Java was designed to be verbose and explicit — which enterprises love because it makes large codebases maintainable.

Who uses it: Google (Android, parts of Search), Amazon (massive Java backend), LinkedIn, Netflix (microservices), Twitter (originally Java), virtually every bank and insurance company, government systems. Java is the language of enterprise at scale.

VS Code extensions:

Extension Pack for Java (Microsoft) — language support, debugging, Maven/Gradle
Spring Boot Extension Pack — Spring development support
Gradle for Java — Gradle build integration
Java Test Runner — run JUnit tests directly in VS Code

Choose it when: You are building Android apps, working in enterprise environments, or targeting jobs at large corporations. Java is verbose compared to Python or Ruby, but its job market is enormous and the pay is high. The learning curve pays off in salary.

Legacy Note There are Java codebases at major banks that are 30 years old. Nobody wants to touch them and the people who know how to maintain them are retiring. This is a money problem that creates a career opportunity for anyone willing to learn Java enterprise patterns.
Data & APIs

Storage and Communication

Every app that does anything real has a database behind it and an API in front of it. These are the fundamentals you need to understand regardless of what language you use.

11

SQL / NoSQL

What SQL is: Structured Query Language — a standard language for interacting with relational databases. In a relational database, data is organized into tables with rows and columns, and relationships between tables are defined by foreign keys. SQL databases include PostgreSQL, MySQL, SQLite, and Microsoft SQL Server.

What NoSQL is: A category of databases that do not use the traditional table-row model. Data is stored in documents (MongoDB), key-value pairs (Redis), wide columns (Cassandra), or graphs (Neo4j). NoSQL databases are flexible — you can change the data structure without a migration.

When to use SQL: When your data has a clear structure and relationships. Financial records, user accounts, orders, inventory. When you need transactions (all-or-nothing operations). When you need complex joins across multiple tables. PostgreSQL is the default right answer for most new projects.

When to use NoSQL: When your data structure changes frequently. When you need to scale horizontally to enormous sizes. When you are storing unstructured content like logs, events, or user-generated content. Firebase Firestore is NoSQL — it is why it handles flexible app data better than a rigid schema would.

VS Code extensions:

SQLTools — connect to and query SQL databases from VS Code
MongoDB for VS Code — browse and query MongoDB directly
Database Client (Weijan Chen) — supports MySQL, PostgreSQL, SQLite, and more
Plain English SQL is like a spreadsheet where every row follows the same format and columns are defined upfront. NoSQL is like a box where you can throw in whatever shape of data you want and find it later. SQL for structure. NoSQL for flexibility.
12

Firebase

What it is: Google's Backend-as-a-Service (BaaS) platform. Firebase gives you a NoSQL database (Firestore), user authentication, file storage, serverless functions, hosting, and analytics — all managed by Google, all connected to a free tier that handles real production traffic before you spend a dollar.

Good for: Solo developers, startups, apps that need to move fast. Firebase removes the need to set up and manage your own server infrastructure. You build the frontend, Firebase handles the backend. Authentication that would take weeks to build securely is a 15-minute setup.

Who uses it: The New York Times, Alibaba, Duolingo, The Economist, Shazam. Firebase is used by 3.5 million apps. Google has strong incentives to keep it running — it feeds developers into the Google Cloud ecosystem.

Core services:

Firestore       — NoSQL document database, real-time sync
Authentication  — email/password, Google, GitHub, Apple sign-in
Storage         — file and image uploads (backed by Google Cloud Storage)
Hosting         — static site and SPA hosting with global CDN
Cloud Functions — serverless backend logic (Node.js or Python)
Analytics       — free event tracking, funnel analysis

VS Code extensions:

Firebase (official) — syntax support for Firestore security rules
Firebase Explorer — browse your Firestore data from VS Code

Choose it when: You are building solo or with a small team, you want to ship fast, and you do not want to manage infrastructure. Firebase free tier handles more traffic than most people expect. You only pay when you scale. It is Billy's primary stack for everything.

Billy's Take Every app Stacked Alchemist has shipped uses Firebase. Authentication, database, hosting, functions — all in one platform, all with a free tier that is genuinely usable. It is not perfect. The pricing model can surprise you at scale. But for getting something live and real? Nothing moves faster.
13

REST APIs

What it is: Representational State Transfer — an architectural style for building web APIs. A REST API is a way for your frontend to communicate with a backend (or a third-party service) over HTTP using standard methods: GET (read), POST (create), PUT/PATCH (update), DELETE (remove).

Why it matters: Every third-party service you integrate — Stripe for payments, Twilio for SMS, SendGrid for email, weather data, sports data, any AI API — communicates via REST. Understanding REST is not optional. It is the plumbing of the modern web.

How a REST request works:

// GET request — fetch a user
GET /api/users/123
Response: { "id": 123, "name": "Billy", "email": "..." }

// POST request — create a new post
POST /api/posts
Body: { "title": "My Post", "content": "..." }
Response: { "id": 456, "title": "My Post", ... }

// DELETE request — remove a record
DELETE /api/posts/456
Response: { "success": true }

Key concepts: Status codes (200 = success, 404 = not found, 401 = unauthorized, 500 = server error), headers (Authorization for API keys, Content-Type for telling the server what format you are sending), and JSON (the data format almost every REST API uses).

VS Code extensions:

Thunder Client — test REST endpoints directly in VS Code
REST Client — send HTTP requests from .http files in your project
Postman (separate app) — the industry standard for API testing and documentation
Plain English A REST API is like a waiter. Your frontend is the customer. The database is the kitchen. You tell the waiter what you want (GET, POST, DELETE), the waiter goes to the kitchen (database/server), and brings back what you asked for. You never talk to the kitchen directly.
14

GraphQL

What it is: A query language for APIs developed by Meta (Facebook). Instead of hitting multiple REST endpoints to gather related data, GraphQL lets the client request exactly the fields it needs in a single query. The server returns exactly that — nothing more, nothing less.

Good for: Complex apps where the frontend needs different shapes of data depending on context. Mobile apps where bandwidth matters (you do not want to fetch 50 fields when you only need 5). Apps with deeply nested relational data that would require multiple REST calls.

Who uses it: GitHub (all public API v4 is GraphQL), Shopify, Twitter/X, Airbnb, PayPal. Meta invented it and uses it across Facebook, Instagram, and WhatsApp at enormous scale. Most teams adopting GraphQL do it because their REST APIs got too complex to maintain.

What a GraphQL query looks like:

// Ask for exactly what you need
query {
  user(id: "123") {
    name
    email
    posts {
      title
      publishedAt
    }
  }
}

// You get back exactly what you asked for
{
  "user": {
    "name": "Billy",
    "email": "...",
    "posts": [{ "title": "My Post", "publishedAt": "..." }]
  }
}

Choose it when: Your REST API has grown into a maze of endpoints and your frontend teams are constantly asking for new ones. GraphQL shifts the power to the frontend — they ask for what they need without waiting for a backend developer to build a new endpoint. It is not the right starting point — build REST first, switch to GraphQL when you feel the REST pain.

💾

Legacy Is Not Dead — It Pays

COBOL still runs the core systems of every major bank in the United States. FORTRAN still powers scientific computing at NASA and CERN. The people who know these languages are retiring. The companies that depend on them are not going anywhere.

Knowing why legacy stacks exist, how they were built, and why companies have not replaced them is real-world knowledge. It is also a career path with almost no competition and extremely high pay. The developer who can walk into a bank and work with 40-year-old COBOL code is not replaceable by a bootcamp grad or an AI. Not yet.

The Point

You Do Not Need to Master All of This

The goal of this page is recognition, not memorization.

Recognition is the real skill

When someone says "we're a Node/Postgres shop" you should know what that means. When a job post says "React + TypeScript" you should know what you are getting into. You do not need to have memorized every API — you need to know what each technology is for and when it is the right tool.

AI fills the syntax gaps

Claude Code can write Python, Java, C#, Rails, or GraphQL. What it cannot do is understand the tradeoff between a relational database and a document store — or know that your company's legacy COBOL system means the new API must speak a specific protocol. That context comes from you. The code comes from AI.

Document what you build

Use AI to generate architecture docs, stack explanations, and deployment runbooks in plain English. Hand those to a client or an interviewer. You will look like you have been doing this for a decade. Because you understand what was built, you just used AI to explain it clearly.

Ready to Start Building?

Pick a stack. Download the CLAUDE.md template. Tell Claude what you are building. Go.