AI Coding Terms for Beginners: A Plain-English Glossary
TL;DR
The words are the hardest part at first. Here is the jargon of AI coding translated into plain English - one clear line each - so the tutorials, tools, and other guides stop being a wall of unfamiliar terms.
The AI words
- Model - the AI itself. The thing that reads your request and produces a response. Different models are better at different jobs.
- Prompt - what you type to the AI. Your instruction or question.
- Token - the small chunks of text an AI reads and writes in. Roughly a few characters each; you are usually billed by how many you use.
- Context window - how much the AI can 'hold in mind' at once. When a conversation gets long, it can run out of room, which is why clearing or summarizing helps.
- Agent - an AI that does not just answer but takes actions in a loop - reading, editing, running commands - to complete a task.
- Hallucination - when an AI states something confidently that is not true. Why you check important output rather than trusting it blindly.
The coding words
- Repository (repo) - a project folder tracked by Git, holding your code and its history.
- Commit - a saved snapshot of your code at a point in time.
- Terminal / command line - the text window where you type commands to your computer.
- IDE / editor - the program you write code in (for example VS Code or Cursor).
- API - a way for one program to talk to another. How your app can use an outside service.
- Deploy - to put your finished project online so other people can use it.
- Debug - to find and fix what is broken in your code.
- Framework / library - prewritten code you build on top of, so you do not start from scratch.
The Claude Code words
- Claude Code - an AI agent that works inside your real project, editing files and running commands with you.
- CLAUDE.md - a file where you write standing facts about your project so Claude remembers them across sessions.
- Slash command - a command you type starting with / to control the session (like /clear or /help).
- Permission mode - how much Claude asks before making changes; you cycle it with Shift+Tab.
- MCP - a standard way to connect an AI tool to outside data and services. Covered in its own guide.
Common questions
What is the difference between a model and an agent?
A model is the AI itself - it reads a request and produces a response. An agent wraps a model in a loop so it can take actions: read files, run commands, check results, and repeat until a task is done. Claude Code is an agent built on a model.
What is a token and why does it matter?
Tokens are the small text chunks an AI reads and writes in - a few characters each. They matter because you are generally billed by how many you use, and because a model can only hold so many at once (its context window). Leaner prompts use fewer tokens.
What does 'deploy' actually mean?
Deploying means putting your finished project somewhere online so other people can use it, rather than it only running on your own computer. It is the step that turns a thing you built into a thing others can visit or use.
Do I need to know all of these before I start?
No. Skim the list once so the words are not a total surprise, then start building and glance back when something confuses you. You learn these terms far faster by using them in context than by trying to memorize them up front.
Get the full plain-English glossary
Get 650+ plug-and-play skills, MCPs & prompts, plus 5,000+ builders - $9/mo, cancel anytime.
Join the Club