30 Claude Concepts Everyone Should Know

Claude keeps adding features, and each comes with its own name. Knowing what each word means helps you pick the right feature and write better prompts.

You do not need to be technical. You just need to know what each button is for.

1 · Model

The AI brain behind Claude; each model has different speed, cost, and skill

2 · Fable, Opus, Sonnet, Haiku

The model families, from most capable to fastest and cheapest; Opus 5.5, Sonnet 5.5, and Haiku 5.5 are current

3 · Mythos

Anthropic’s most capable model line, offered with limited availability, including for security work

4 · Token

A small chunk of text; usage and limits are counted in tokens

5 · Context window

How much Claude can read at once; up to 1 million tokens on recent models in some products

6 · System prompt

Standing instructions that shape every answer

7 · Extended thinking

Claude reasons step by step before answering hard questions

8 · Hallucination

A confident answer that is wrong; always check facts that matter

9 · Projects

Workspaces with their own files and instructions

10 · Artifacts

Documents, code, and apps Claude makes in a side panel

11 · Memory

Claude remembering useful details across chats, which you can view and turn off

12 · Styles

Preset or custom writing styles for responses

13 · Research

Claude searching many sources and writing a cited report

14 · Connectors

Links to apps such as Gmail, Google Drive, Slack, and Notion

15 · MCP

Model Context Protocol, the open standard that connectors use

16 · Skills

Folders of instructions and files that teach Claude a task, loaded only when needed

17 · Plugins

Bundles of skills, connectors, and commands you install in one go

18 · Claude Code

Claude as a coding agent in your terminal, editor, desktop app, or browser

19 · CLAUDE.md

A file of project rules Claude reads at the start of every session

20 · Plan mode

Claude plans first and waits for your approval before changing files

21 · Hooks

Scripts that run automatically at set moments, such as after an edit

22 · Subagents

Helper agents Claude sends off to handle parts of a task

23 · Cowork

Claude working on multi-step tasks with your files and apps on your desktop

24 · Claude in Chrome

Claude reading and acting on web pages in your browser

25 · Claude Design

Claude making visual designs, decks, and prototypes

26 · Motion and Dashboards

Animated explainers and live dashboards made from your content

27 · API

How developers build Claude into their own apps

28 · Prompt caching

Reusing a long prompt cheaply across many requests

29 · Managed Agents

Agents Anthropic hosts and runs for developers

30 · Usage limits and the constitution

Plans set how much you can use Claude; Claude’s constitution sets the values it follows

Features and plan access change often. Check Anthropic’s help centre for what your plan includes today.

• Use a Project for anything you return to
• Turn on extended thinking for hard problems
• Add connectors for the apps you use daily
• Write a CLAUDE.md before your first coding session
• Pick the smallest model that does the job well