This guide breaks down common terms you’ll run into when working with code and AI tools. We’ve kept the definitions simple so you can get up to speed quickly.
Atoms-specific terms
Team Mode: Activates a team of AI agents to handle complex requirements for you, including a Product Manager, Engineer, Data Scientist, Deep Researcher, and Architect.
Engineer Mode: Assigns only the Engineer agent, Alex, to work on your project.
Race Mode: Allows you to run the same prompt across multiple models simultaneously to compare their outputs.
Deep Research: A feature for conducting deep research to generate professional reports, including academic papers, business analyses, and market reports.
App Viewer: Allows you to preview results and modify any visual element directly on the page.
Remix: Lets you duplicate any project into your own chat to continue building or editing from there.
Publish: Generates a stable link to your current application for easy sharing.
App World: Atoms' platform for creativity, where you can display your work and highlight your skills on the homepage.
LLMs & Vibe Coding
Vibe Coding: The practice of writing software by telling an AI what you want in plain English (the "vibe" or intent) and letting it generate the actual code. You focus on whether the app feels right and works correctly, rather than worrying about semicolons or specific logic.
Artificial Intelligence (AI): A broad term for computer systems that can do tasks that usually require human intelligence—like recognizing speech, making decisions, or translating languages.
LLM (Large Language Model): A type of AI trained on massive amounts of text. It's the technology behind tools like ChatGPT and Claude. It’s incredibly good at predicting what word comes next in a sentence, which allows it to write essays, answer questions, and write code.
Hallucination: When an AI confidently gives you a completely wrong answer. AI models are great at sounding convincing, so you should always double-check the code or facts they generate.
Prompt: The text you type to the AI to get an answer. "Prompt Engineering" is just a fancy way of saying "learning how to ask the right questions to get the best results."
Token: How an AI counts. An AI doesn't see "words"—it sees "tokens." A token is roughly four characters of text. For example, the word "apple" is one token, but a long, complex word might be three tokens. AI costs and limits are usually measured in tokens.
Human-in-the-Loop: That’s you. Even though the AI is writing the code, you are the "manager" in the loop. You review the work, spot checks, and guide the AI if it goes off track. You stop being the writer and become the reviewer.
Iterating: The process of refining the code through conversation. In vibe coding, you rarely get it perfect on the first try. You "iterate" by saying things like, "Make that button bigger," "Fix the error on line 10," or "No, use a different color."
Coding Basics
API (Application Programming Interface): A bridge that lets two different programs talk to each other. Think of it like a waiter in a restaurant: you (the user) give an order to the waiter (the API), and they bring you the food from the kitchen (the system) without you needing to know how it was cooked.
Bug: A mistake or glitch in the code that causes the software to crash or behave strangely.
Version Control (Git): A "time machine" for code. It tracks every change anyone makes, so if something breaks, you can easily go back to a version that worked.
Frontend (Client-side): This is everything you can see and touch on your screen. It’s the buttons, the text, the images, and the animations. If you are browsing a website, the "frontend" is what is running in your browser.
Backend (Server-side): This is the "brain" of the application that lives on a remote computer (server). It processes data, saves your files, and manages the logic that the user never sees.
UI (User Interface): The specific design and layout of the app. It refers to the "look and feel"—the color schemes, the shape of the buttons, and the readability of the text.
Full Stack: A developer (or a tool) that can handle both the Frontend and the Backend.
Software Development Concepts
Product Manager: The person who decides what to build and why. They act like the "CEO" of a specific feature, balancing what the users want with what the business needs. They write the plans (PRD) but usually don't write the code.
Engineer: The builder. They take the plans from the Product Manager and write the actual code to make it work. They figure out how to build it efficiently and securely.
Architect: The "city planner" of the software. They don't usually write every line of code; instead, they make the high-level decisions about which technologies to use and how different parts of the system should connect so the whole thing doesn't collapse as it grows.
Data Scientist: The detective. They look at the massive amounts of data the app generates (like user clicks or sales numbers) and use math and code to find patterns, helping the business make smarter decisions.
Deep Researcher: A role (or sometimes a specialized AI agent) focused on investigating complex, unknown topics. Unlike a standard search, a deep researcher digs into academic papers, technical documentation, and long reports to synthesize a comprehensive answer on a difficult subject.
Strategy & Growth
SEO (Search Engine Optimization): The art of tweaking your website so it shows up at the top of Google results. It involves using the right keywords, making the site fast, and ensuring other sites link to you.:
Growth: A mix of marketing and engineering focused entirely on getting more users. A "Growth Engineer" might build features specifically to encourage people to invite their friends or sign up faster.
PRD (Product Requirements Document): The "blueprint" for a feature. Written by the Product Manager, this document explains exactly what a new feature should do, how it should look, and how success will be measured. Engineers use this as their instruction manual.
Page structure & Navigation
Headings (Titles): The outline of your page content. In code, these are marked as H1, H2, H3, etc. H1 is the main title (like a book title), and H2s are the chapter titles. They help both users and Google understand what your page is about.
Footer: The section at the very bottom of every page. It’s the "junk drawer" of the website—where you put things that are important but not the main focus, like legal policies, contact info, and links to social media.
Favicon: The tiny icon that appears on the tab of your web browser (next to the page title). It helps users identify your site quickly when they have twenty tabs open at once.