This is where you'll live for the rest of the course. So let's walk the rooms.

Claude is built by Anthropic, a company founded in 2021 by former OpenAI researchers. That matters — not because of the corporate history, but because of the design philosophy underneath it. They built Claude around a principle called Constitutional AI, which in plain terms means: the tool is trained to be helpful and honest before it's trained to be impressive.

You feel that difference immediately when you start writing with it. Claude doesn't try to dazzle you. It tries to understand you. And for creative work — where the voice has to stay yours — that distinction changes everything.

When you open Claude, you're not using one AI. You're choosing from a family of models, each built for a different kind of work. Think of it like engines — same manufacturer, different horsepower.

There's a feature called Extended Thinking that changes how Claude processes complex requests. When it's on, Claude doesn't just generate a response — it reasons through the problem first. You'll see a thinking indicator while it works, and the output that comes out the other side is noticeably more considered.

Think of it as the difference between asking someone a question while they're walking past your desk versus sitting down with them and saying "take your time with this." Same person, same brain — but the quality of attention is different.

Extended Thinking is especially useful for structural decisions: outlining a book, evaluating a complicated draft, or working through plot logic. It's not a separate model. It's a mode you can toggle on when the moment calls for it.

If Claude is the office, Projects are the rooms inside it.

A Project is a dedicated workspace you create around a specific task, goal, or body of work. It holds its own knowledge base — documents, files, reference material you upload — and its own custom instructions that apply to every conversation inside it. Start a new chat in that Project, and Claude already knows the context. No re-explaining. No warm-up.

This course was built inside a Claude Project. Every module, every design decision, every editorial choice — it all lives in one workspace. Claude knows the voice, the audience, the structure, the terminology. When we start a new conversation in the Project, it picks up exactly where the work left off.

The knowledge base holds up to roughly 500 pages of material. Upload your manuscript draft, your research notes, your outline, your style guide — whatever the AI needs to stay informed. It references that material automatically across every conversation in the Project. You don't have to keep pasting things in.

Free users can create up to five Projects with basic functionality. Pro and Max plans unlock the full feature set — larger knowledge bases, more Projects, and the ability to share them with collaborators on Team plans.

Inside every Project, there's a space for custom instructions. This is the back channel — the set of standing orders Claude follows in every conversation within that workspace. The AI reads them before it reads your first message. They shape everything that follows.

Think of it like briefing a new team member on day one. You're not giving them a task yet. You're telling them how you work: "We write in short paragraphs. We don't use bullet points. The tone is direct but warm. The audience is 55 and older. When I say 'Fire Mode,' I want unfiltered critical feedback."

Those instructions persist. You set them once, and every conversation in that Project inherits them. No repeating yourself. No re-establishing context. The AI arrives pre-calibrated.

The power here is compounding. As you refine those instructions over time — adding new preferences, clarifying what works and what doesn't — the AI gets sharper. Not smarter in a technical sense, but more aligned with how you think and what you expect. The instructions become a living document that evolves with your working relationship.

We'll build your first set of custom instructions together in a later module. For now, just know the space exists. It's waiting.

Claude has a memory system that works across conversations. It remembers things about you — your preferences, your working patterns, key details from past sessions — and carries that knowledge forward into new conversations without you asking.

This is separate from Projects. Memory is account-wide on paid plans. It synthesizes what it learns from your conversations every 24 hours and updates its understanding of who you are and how you work. Each Project also builds its own separate memory, focused on that specific body of work.

You control it. You can see what Claude remembers, edit it, pause it, or reset it entirely. If you want a conversation that doesn't touch memory at all, there's an incognito mode — the ghost icon in the upper right corner of your screen.

There's a feature called Styles that lets you adjust how Claude communicates — separate from what it knows. You can switch between preset styles or create your own based on your writing. Concise for quick answers. Explanatory when you're learning. A custom style built from samples of your own prose.

It's the difference between telling someone what to say and telling them how to say it. Styles handle the how. Custom instructions handle the what. Together, they shape an AI that sounds less like a machine and more like a collaborator who knows your voice.

When Claude creates something substantial — a document, a code file, an interactive component — it appears in a dedicated window alongside the conversation. These are called Artifacts. You see the work take shape in real time while the conversation about the work continues in the chat.

This matters for creative writers more than you'd expect. You can ask Claude to draft a chapter outline, and it appears as a living document you can review, edit, and iterate on while you're still talking through ideas. No copying and pasting between windows. The work and the conversation about the work live side by side.

Claude can also create files directly — Word documents, presentations, spreadsheets, PDFs. Ask for it, and it builds it. The modules you're reading right now? Built as HTML files, produced right inside the conversation, then deployed to a live website. All from within Claude.

Claude can search the web in real time. If you ask about something current — a recent event, today's pricing, a fact that might have changed since the AI's training data was locked — it goes and finds it. No guessing. No outdated information dressed up as confidence.

Beyond search, Claude connects to external tools through integrations. Google Drive. Notion. The list keeps expanding. These connections let Claude pull from your existing files and workflows, not just its own knowledge. Your research, your notes, your existing documents — accessible right inside the conversation.

For writers, this means your AI partner can fact-check itself, pull from your reference material, and stay current — without you having to open a separate browser tab and do it manually.

Here's what it looks like when all of this works together.

You create a Project for your book. You upload your draft, your character notes, your research. You write custom instructions that describe your voice, your audience, and your expectations. Claude's memory already knows your working style from previous sessions. You've set a Style that matches your prose.

You open a new conversation in that Project. You type one sentence — "Let's work on Chapter 7." And Claude arrives already knowing who you are, what you're building, how you want it to sound, and where you left off. No preamble. No setup. Just the work.

You don't need to set all of this up today. We'll build it together, piece by piece, across the modules ahead. This tour was just to show you the rooms — so when we start moving furniture in, you'll know where everything goes.