AI Open Guide

All the words the guide explains, in one list, with the chapter where each one first appears.

Account Chapter 0.2
Your login; one account works in every door to Claude
Action Chapter 2.10
Anything that changes the world outside the chat: saving, sending, clicking, moving
Add-in Chapter 0.2
A Claude panel installed inside Excel, Word or PowerPoint
Agent Chapter 6.8
Claude working on its own through several steps, with tools and a goal
Agent loop Chapter 6.8
Decide, use a tool, read the result, decide again, until done
Agent SDK Chapter 6.8
A library that puts the Claude Code engine inside your own program
AI Chapter 0.1
A program that learned from a huge amount of text how to understand and write language
Allow once / Always allow Chapter 2.3
The choice you get the first time Claude wants to use a website
API Chapter 6.6
A door for programs: send a request, get an answer back as data
API key Chapter 6.6
The long secret string that identifies your account to the API; treat it like a password
App Chapter 0.2
A program installed on your computer or phone
Artboard Chapter 3.1
One frame on the Design canvas, such as "Landing page — desktop"
Artifact Chapter 1.8
Anything Claude makes that lives beside the chat: a document, deck, sheet, design, app
Attach / Upload Chapter 1.2
Putting a file into the chat so Claude can read it
Batch Chapter 6.7
Many API requests answered within hours, at half price
Brainstorm Chapter 0.4
Produce many ideas quickly without judging them yet
Brand Chapter 3.2
The consistent look and voice of a company or person
Brief Chapter 4.1
A short, clear description of what you want, for whom, and what good looks like
Browser Chapter 2.3
The program that shows websites; Claude has a built-in one and can use Chrome
Browser extension Chapter 0.2
A small add-on inside Chrome; Claude in Chrome is one
Bundle Chapter 3.5
The package a Design handoff creates for Claude Code
Canvas Chapter 3.1
The area in Design where the picture appears
Chat Chapter 0.1
One conversation with Claude
Clarifying question Chapter 4.4
A question Claude asks to fill a gap in your request
CLAUDE.md Chapter 6.2
A text file in a project with the rules Claude Code reads at the start of every session
Claude Code Chapter 6.1
Claude working inside your files and programs: reads, changes and runs code
Client Chapter 6.5
The side that uses MCP tools (Claude Code, the app, your program)
Cloud / Cloud session Chapter 1.12, 2.1
Running on Anthropic's computers, so your laptop can be closed
Code Chapter 6.1
The text instructions that make software work
Computer use Chapter 2.4
Claude sees your screen and uses your mouse and keyboard
Confidence Chapter 4.8
How sure Claude is; ask and it will tell you
Connect a folder Chapter 2.2
Choosing one folder on your computer that Claude may use
Connector Chapter 1.9
A bridge that lets Claude read or act inside another app, like Gmail
Console Chapter 6.6
Anthropic's website for developers: keys, usage, spending limits
Constraint Chapter 4.2
A rule the answer must follow: length, language, style, things to avoid
Context Chapter 0.5
Everything Claude can see right now: your message, files, instructions, the chat so far (the desk)
Context window Chapter 0.5
The size of the desk; big but not endless
Convention Chapter 6.2
A habit a project follows: naming, language, how to run tests
Cowork Chapter 2.1
The "do it" side of Claude: plans, uses tools, makes files; now part of every chat, not a separate mode
Cron Chapter 6.9
The old way of writing a schedule like "every Monday at 9"
Deck Chapter 3.3
A set of slides for a talk or pitch
Deletion protection Chapter 2.10
Claude needs a clear Allow before permanently deleting a file
Deliver Chapter 1.12
Where a scheduled task's result goes
Design Chapter 3.1
A visual layout: picture, headline, buttons, colors
Design system Chapter 3.2
Saved rules for how your things should look: colors, fonts, spacing, buttons
Desktop app Chapter 0.2
The Claude app for your computer; folder access and computer use live there
Dictation Chapter 1.10
You speak, Claude turns it into text in the box
Draft Chapter 0.4, 4.5
A first version, good enough to react to
Effort Chapter 1.1
How hard the model thinks before answering: low, medium, high
Example Chapter 4.3
A real sample of what you want; teaches more than a description
Export Chapter 1.8
Save an artifact as a file: Word, PDF, PowerPoint, or send to Google Docs or Notion
Extract Chapter 1.2
Pull specific things out of a document: dates, amounts, names
Folder instructions Chapter 2.2
A note Claude reads whenever it works in that folder
Format Chapter 4.2
The shape of the answer: list, table, email, file
Formula Chapter 2.6
A spreadsheet cell that calculates and updates when numbers change
Generate memory from chats Chapter 1.6
The switch that turns memory on; on by default on Free, Pro and Max
Generic Chapter 4.7
An answer that could have been written for anyone
Grant an app Chapter 2.4
Choosing which programs computer use may control
Guessing Chapter 4.4
What Claude does when a gap in your request is left open
Hallucination Chapter 0.5
When Claude writes something that sounds right but is made up
Handoff Chapter 3.5
Passing a design from Claude Design to Claude Code to build
Helper (sub-agent) Chapter 2.5, 6.4
A separate Claude that handles one part of a task and reports back
Hook Chapter 6.4
A rule in Claude Code: when X happens, run Y, automatically
In parallel Chapter 2.5
At the same time instead of one after another
Instructions for Claude Chapter 1.7
The Settings box where you write how Claude should behave everywhere
Iteration Chapter 4.5
Doing something again, a little better each time
Kind of ask (pattern) Chapter 4.6
The type of job: create, analyze, research, transform, automate, build, decide, learn
Knowledge cutoff Chapter 0.5
The date Claude's built-in knowledge ends; newer facts need search
Lifecycle event Chapter 6.4
A moment a hook can attach to: session start, before or after a tool, finish
Logged-in session Chapter 2.3
Being signed in to a site in the browser, so Claude can act as you there
Managed Agents Chapter 6.8
Anthropic hosts your agent: loop, sandbox, memory, scheduling
Master Prompt Chapter 4.2
A request with all nine blocks, in a fixed order
MCP Chapter 6.5
The open standard for how an AI talks to tools; every connector is MCP underneath
Memory Chapter 0.5, 1.6
The small notebook of facts Claude keeps about you between chats
Mockup Chapter 3.1
A picture of what something would look like before it is built
Model Chapter 0.3, 1.1
A version of Claude's brain: Haiku, Sonnet, Opus, Fable; faster or smarter
Off-topic Chapter 4.7
An answer about something near your question, not your question
One-pager Chapter 3.3
A single designed page that explains one thing
Outputs Chapter 2.6
The Cowork panel listing the files Claude made, with Download
Permission (authorize) Chapter 1.9
The step where another app lets Claude in
Permission mode Chapter 2.10, 6.10
How Claude handles actions: ask every time, check automatically, or skip checks
Plan Chapter 0.3
Your package: Free, Pro, Max, Ultra, Team, Enterprise ("Max 5x" is now Max, "Max 20x" is now Ultra)
Plugin Chapter 1.11
A bundle of skills, connectors and helpers for a role, installed in one click
Progress panel Chapter 1.4
The box on the right during long tasks, listing the steps
Progressive loading Chapter 6.3
Claude reads only a skill's one-line description until it is needed
Project Chapter 1.5
A workspace with its own instructions, files and chats
Project files (knowledge) Chapter 1.5
Documents added to a Project once, readable in every chat there
Project instructions Chapter 1.5
The standing text Claude reads at the start of every chat in a Project
Prompt Chapter 0.1
The message you type to Claude
Prompt caching Chapter 6.7
The API charges much less for long text repeated in every request
Prompt injection Chapter 6.10
Text inside a page or document that tries to give Claude instructions
Prototype Chapter 3.4
A pretend app or site that looks and clicks like the real thing
Pull request (PR) Chapter 6.9
A proposed change to a repo that others review before it is merged
Rate limit Chapter 6.10
How many API requests are accepted per minute
Read and write Chapter 1.9
Some connectors only let Claude look; others let it act
Remote vs local (MCP) Chapter 6.5
A server on the internet vs one installed on your computer
Repository (repo) Chapter 6.1
A project folder under version control, so every change can be undone
Request and response Chapter 6.6
What you send to the API and what comes back
Research Chapter 1.4
A multi-step search: plan, search, compare, verify, write
Reset Chapter 0.3
Usage limits refill on a timer
Review Chapter 4.5
Reading something to find what is wrong before fixing it
Role Chapter 4.2
Who you want Claude to be for a task
Routine Chapter 6.9
A scheduled or triggered job in Claude Code
Sandbox Chapter 6.8
An isolated computer where an agent can run code safely
Scheduled task Chapter 1.12
A job with a timetable that Claude runs without you
Scope Chapter 4.7, 6.2
How wide a question is; also where a rule applies
Screen Chapter 3.4
One view of an app in a prototype
Secret Chapter 6.10
Any key, password or token; never in a prompt, repo or screenshot
Selection Chapter 2.7
The cells or text you have highlighted for the add-in to work on
Sensitive topics Chapter 1.6
Health, religion, politics and similar; a separate memory switch
Server (MCP) Chapter 6.5
The program that offers tools over MCP
Session Chapter 6.1
One working conversation in Claude Code about one folder or repo
Side panel Chapter 2.7, 2.8
The strip beside a document, browser or sheet where you talk to Claude
Skill Chapter 1.11, 6.3
A packaged "how to" for one job that Claude loads when the job matches
SKILL.md Chapter 6.3
The file at the heart of a skill: name, when to use it, the steps
Slack Chapter 2.9
The team chat app; Claude can be tagged in it on Team plans
Slash menu Chapter 1.11
Type / in the chat box to see skills you can call by name
Source Chapter 1.3, 4.8
Where a fact came from; click the tag to open the page
Spending limit Chapter 6.10
A cap set in the console; the API stops when reached
Standing rule Chapter 1.7
Something true every time: "short answer first"
Structured output Chapter 6.7
The API's answer always fits a shape you give it
Style Chapter 4.3
The way something is written; hard to describe, easy to show
Summarize Chapter 0.4
Make something long into something short that keeps the important points
Tab Chapter 2.8
One open web page
Tag (mention) Chapter 2.9
Typing @Claude in Slack so it sees the message
Take a screenshot Chapter 2.4
Claude looks at your screen once; the look-only form of computer use
Task (Cowork) Chapter 2.1
One job from start to finish, with its own session
Team and Enterprise plans Chapter 2.9
The company plans of Claude
Template Chapter 4.6
A request with blanks in it
Terminal (command line, CLI) Chapter 6.1
A text window where you type commands to the computer
Thinking (extended thinking) Chapter 1.1
Claude reasoning step by step before the final answer
Token Chapter 6.6
The unit Claude counts text in, about three-quarters of a word
Tool Chapter 2.1
Something Claude can use during a task: a skill, connector, browser, file
Tool use (API) Chapter 6.7
Claude asks to call functions your program has, with the right inputs
Transcript Chapter 1.10
The written version of what was said
Trigger Chapter 6.3, 6.9
What makes a skill load, or what starts a routine
Typography Chapter 3.2
The fonts and how they are used
Usage limit Chapter 0.3
How much you can ask before Claude tells you to wait
Vague Chapter 4.1
Missing the details that would let someone do the job well
Verify Chapter 4.8
Confirm something is true by looking at a second place
Voice mode Chapter 1.10
A spoken back-and-forth with Claude, mainly on the phone
Web search Chapter 1.3
Claude reads live web pages to answer you
Wireframe Chapter 3.4
A rough prototype in grey boxes
Workflow Chapter 2.5
A bigger, scripted version of helpers: many in planned stages