The AI Build Playbook

A text-based way of giving instructions to your computer by typing commands rather than clicking buttons. It looks intimidating but you only need a handful of commands to get started.

A specific instruction you type into the terminal to make something happen.

An AI-powered code editor that sits on top of your existing development environment. Popular with people building with AI for its ability to understand and edit large codebases through conversation.

The process of finding and fixing errors in your code. An inevitable and normal part of building anything.

The process of taking your app from running locally on your computer and making it live on the internet so anyone can access it via a URL.

An error is a problem that stops your code from running. A bug is unexpected behaviour where the code runs but does not do what you intended.

The most popular version control tool. It tracks changes to your project files and lets you save snapshots of your work called commits.

A website where you store and manage your Git repositories online. Think of it as Google Drive for your code. It also makes it easy to deploy your project to platforms like Vercel.

A software application that gives you everything you need to build in one place. Think of it as your workshop. It typically includes a code editor, file manager and tools to run and test your project.

Running your app on your own computer rather than on the internet. When your app is running locally you access it through your browser using an address like localhost:3000. Nobody else can see it yet.

A lightweight way of formatting text using simple symbols rather than a visual editor. Widely used in documentation, README files and AI outputs. Once you recognise the syntax it becomes second nature.

An open standard developed by Anthropic that allows AI models to connect to external tools, data sources and services in a consistent way. Think of it as a universal plug socket for AI integrations.

A popular JavaScript library for building user interfaces. It lets you build reusable components, pieces of UI that you can combine to create full applications. Most modern web apps built with AI tools use React under the hood.

Improving the structure or quality of existing code without changing what it does. Makes it easier to read, maintain and build on.

A folder that contains all the files for your project, tracked by Git. Every project has its own repository.

The combination of tools, languages and platforms used to build an application.

A utility-first CSS framework that lets you style your application by applying pre-built class names directly to your HTML elements rather than writing custom CSS. Extremely popular in modern web development because it speeds up the design process significantly.

The application on your computer where you type CLI commands. On a Mac it is called Terminal. On Windows it is called Command Prompt or PowerShell.

A system for tracking changes to your code over time so you can see what changed, when and why. It also lets you roll back to a previous version if something breaks.

Building software by describing what you want to AI in plain language, rather than writing every line of code yourself. It is one way of building with AI, and the term you will often hear for it.

AI that does not just respond to questions but takes actions, makes decisions and works through multi-step tasks on your behalf without needing to be told every single step.

The underlying intelligence that powers AI tools. Different models have different strengths, costs and capabilities.

A way for two pieces of software to talk to each other. When your app needs to use an AI model, it sends a request to the model's API and gets a response back.

A unique private code that identifies you when your app makes requests to an API. Treat it like a password. Never share it publicly.

The maximum amount of text an AI model can process in one go. Think of it as the model's short term memory. If your conversation exceeds the context window the model starts to forget earlier parts.

Training an existing AI model further on specific data to make it better at a particular task or more aligned with a specific style or domain.

When an AI model generates information that sounds confident and plausible but is factually incorrect. A known limitation of all current LLMs.

The type of AI model that powers tools like ChatGPT, Claude and Gemini. Trained on vast amounts of text data, they understand and generate human language extremely well.

The instruction or question you give to an AI model. The quality of your prompt directly affects the quality of the output.

A technique that allows an AI to answer questions based on your own documents and data rather than just its general training. The AI retrieves the most relevant parts of your documents before generating a response.

A set of instructions given to an AI model before the conversation begins that shapes how it behaves throughout. Used to give the AI a persona, rules or context.

The unit AI models use to measure text. A token is roughly three to four characters or about three quarters of a word. Models charge by the number of tokens processed.

The process of verifying who a user is before giving them access to an application. Usually handled via email and password, or third party login like Google.

The part of an application that runs behind the scenes. It handles logic, stores data, processes requests and communicates with databases and APIs.

The four fundamental operations you can perform on data in any database driven application. Almost every app you build will use all four.

Where your application stores information persistently. Without a database, data disappears when you close the app.

A file where you store sensitive information like API keys so they are not visible in your code. The dot before env means it is a hidden file.

A structured foundation for building applications that provides common tools, patterns and rules so you do not have to start from scratch.

The part of an application that users see and interact with. Everything visual: buttons, forms, layouts, colours.

Where your application lives on the internet. A hosting provider stores your app's files on a server and makes them accessible via a URL.

A JavaScript runtime that lets you run JavaScript outside of a browser, on your computer or a server. Many modern web tools require Node.js to be installed.

A tool that lets you install and manage packages, which are pre-built chunks of code that add functionality to your project. It comes with Node.js.

Software whose code is publicly available for anyone to view, use and modify.

Pre-built code that someone else has written and shared that you can add to your project to save time.

A way of running backend code without managing your own server. The hosting provider handles all the infrastructure and you only pay for what you use. Scales automatically with demand.

An open source database platform that is beginner friendly and works well with apps built with AI. It handles user authentication, data storage and real time updates.

A popular deployment platform that makes it very easy to take a project from GitHub and make it live on the internet. Free tier is generous for small projects.

A way for one application to automatically notify another when something happens. Rather than constantly checking for updates, the app sends a message the moment an event occurs.

The rate at which subscribers cancel their subscription. Keeping churn low is as important as acquiring new customers.

A small, focused software product built by one person or a tiny team, usually solving one specific problem for a specific audience. Sold on a subscription basis.

The predictable revenue your product generates each month from subscriptions.

The simplest version of your product that is functional enough to test with real users and get feedback from.

An app that is stable, secure and reliable enough to be used by real people in the real world, not just a demo or prototype.

Software delivered over the internet on a subscription basis rather than as a one-time purchase. Users access it through a browser without installing anything.

Shorthand for version numbers. v1 is the first working version. v2 is the improved version after feedback. Iteration is how great products are built.

A list of people who have expressed interest in your product before it launches. Building a waitlist before you build validates demand and gives you an audience ready to convert.