A

Artifacts

Standalone pieces of content generated by the AI - like a full HTML file, a React component, or a Markdown document. Useful because you can copy them whole instead of picking apart a chat response.

Agentic AI

AI systems that don't just answer questions, but can take actions, use tools (like writing files or searching the web), and follow a plan to achieve a goal autonomously.

C

Context Window

The maximum amount of text (tokens) an AI can process in one go. If your prompt is too long or you've been chatting too much, the AI 'forgets' earlier instructions and output quality drops drastically.

CSS Variables (Custom Properties)

Tokens defined in CSS (e.g., --forest: #1C3D2E) that allow you to change colors or spacing globally across an entire site simply by updating one value.

F

Few-Shot Prompting

Providing the AI with a few examples of what you want before asking it to do the task. For example, showing it a button design before asking it to build a full form. Always better than Zero-Shot.

H

Hallucination

When the AI confidently makes something up. In coding, this often looks like using CSS classes that don't exist in your framework, or calling functions that haven't been written.

P

Prompt Chaining

Breaking a complex task into smaller, sequential prompts. Instead of 'build a website', you ask it to 'build the HTML structure', then 'add CSS styles', then 'add interactions'.

T

Token Limit

Similar to context window, but specifically refers to the hard limit on the number of tokens (words/pieces of words) the AI can read or generate in a single request.

V

Vibe Coding

The act of using natural language to direct AI agents to build software for you. You provide the 'vibe' (the design intent, the logic flow), and the AI handles the syntax.

Z

Zero-Shot Prompting

Asking the AI to do a task without giving it any examples. It relies entirely on its pre-trained knowledge. Good for simple tasks, risky for complex designs.