Install Claude Desktop on your machine, understand the 6 capabilities the web app doesn't have, configure your account, and complete a real work task that would have taken you 20 minutes manually.
Install Claude Desktop on your machine, understand the 6 capabilities the web app doesn't have, configure your account, and complete a real work task that would have taken you 20 minutes manually.
Claude Desktop is available for macOS and Windows. The installation is straightforward — the interesting part comes after.
Go to claude.ai/download and download the version for your operating system. macOS requires 12+ (Monterey or later). Windows requires Windows 10 or later.
On macOS, drag Claude to your Applications folder. On Windows, run the installer. Sign in with your Anthropic account. If you don't have one, create a free account at claude.ai first — the Pro subscription ($20/month) unlocks higher usage limits and is worth it if you plan to use MCP seriously.
Free vs Pro for Desktop: The free tier works for everything in this course. Pro gives you 5x the usage and access to Claude's most capable models. For the exercises in Day 3-5 (MCP), the free tier is fine. You can always upgrade later.
These will save you time every single day:
| Action | macOS | Windows |
|---|---|---|
| New conversation | ⌘ N | Ctrl N |
| Open from anywhere (global) | ⌘ Space then "Claude" | System tray icon |
| Submit message | Return | Enter |
| New line in message | Shift Return | Shift Enter |
| Navigate conversations | ⌘ K | Ctrl K |
| Copy last response | Click the copy icon | Click the copy icon |
The web interface is a chatbot. The desktop app is something different. Here's exactly what you get that the web can't do:
| Capability | Web (claude.ai) | Desktop App |
|---|---|---|
| Chat with Claude | ✓ Yes | ✓ Yes |
| File uploads (manual) | ✓ Yes | ✓ Yes |
| Projects with persistent context | ✓ Yes | ✓ Yes |
| MCP (connect to your computer) | ✗ No | ✓ Yes |
| Read files without uploading | ✗ No | ✓ Via MCP |
| Run code / shell commands | ✗ No | ✓ Via MCP |
| Connect to databases | ✗ No | ✓ Via MCP |
| Connect to GitHub / repos | ✗ No | ✓ Via MCP |
| Global keyboard shortcut | ✗ No | ✓ Yes |
| Works offline (cached) | ✗ No | ✓ Partial |
MCP is the unlock. We cover it fully on Day 3. For now, understand that MCP lets Claude act on your computer — reading files, querying databases, pushing code — rather than just talking about them.
What is MCP exactly? MCP stands for Model Context Protocol. In plain English: it's a standard way for Claude to connect to small background programs (called "servers") that give Claude new abilities. Each server adds a specific capability. The filesystem server lets Claude read files on your computer without you uploading them. The GitHub server lets Claude look at and modify your code repositories. The database server lets Claude run queries against your data. You install a server once and it runs quietly in the background — Claude can use it whenever it's relevant.
Don't ask Claude "what can you do?" That's a waste of your first conversation. Instead, give it a real task you'd normally do manually. Here's a good one to start:
Find the last work email you drafted that felt too long or unclear. Paste it into Claude with this prompt:
Rewrite this email to be clearer and more direct. Keep it under 150 words. Use plain language. The recipient is [describe who they are]. [paste your email here]
Notice what happens: Claude produces a rewrite immediately, and you can iterate. "Make it more formal." "Add a call to action at the end." "Shorter." Each instruction refines the output. This back-and-forth is the core interaction pattern you'll use everywhere.
Every Claude Desktop conversation follows this structure:
The mistake most people make is stopping at step 2. The real value comes from iteration. Claude doesn't get tired. It doesn't take your edits personally. Use it.
Pro tip: Start every new work session with "This lesson is working on [X] today." Claude can't remember your previous conversations (by default), so a short context-setter at the start saves you from re-explaining things mid-conversation.
Your conversation history lives here. Use ⌘ K (Mac) or Ctrl K (Windows) to search across all your past conversations — this is genuinely useful once you've been using Claude for a few weeks and need to find something you worked on.
Claude Desktop gives you access to multiple models. For daily work, Claude Sonnet is the best balance of speed and capability. Use Claude Opus for complex reasoning tasks where you can afford to wait 10-15 seconds longer for a response. You'll find the model selector in the conversation toolbar.
When Claude produces structured output — a chunk of code, a long document, a table — it often places it in a separate panel called an "artifact." Think of it as a side-by-side view: your conversation on the left, the created content on the right. You can edit artifacts directly in that panel, ask Claude to update them, or copy and export them. We cover artifacts in depth on Day 4.
Don't just read this course. Do the work.
Tomorrow: Projects. You'll build a persistent knowledge base that makes Claude an expert in your specific work domain.
The foundations from today carry directly into Day 2. In the next session the focus shifts to Day 2 — building directly on everything covered here.
Before moving on, verify you can answer these without looking:
Live Bootcamp
Learn this in person — 2 days, 5 cities
Thu–Fri sessions in Denver, Los Angeles, New York, Chicago, and Dallas. $1,490 per seat. June–October 2026.
Reserve Your Seat →