Understanding the Interface
Module: Getting Started | Lesson: 6 of 6 | Time: ~10 minutes
What You Will Learn
- How the Claude Code screen is laid out
- How permission prompts work and how to respond to them
- What the different tool indicators mean
- Keyboard shortcuts for the input area
- Useful slash commands you can use while chatting
- How to toggle verbose mode for more detail
Prerequisites
- Completed Your First Conversation
The Claude Code Interface
Claude Code is a text-based interface. There are no buttons, menus, or toolbars — you interact entirely by typing. If you have used a chat application before, the layout will feel familiar:
- Claude's responses appear at the top of the screen, scrolling upward as the conversation grows
- Your input area sits at the bottom, where you type your messages
That's it. Simple by design. The power comes from what you type, not from clicking around.
If you are used to graphical applications with lots of buttons, a text-only interface might feel strange at first. Give it time — most people find it faster once they get used to it.
Permission Prompts
This is one of the most important things to understand about Claude Code. When Claude wants to do something on your computer — create a file, run a command, edit code — it asks you first. It cannot act without your approval.
You will see a prompt that looks something like this:
Claude wants to use the Write tool to create hello.txt
Allow? [Y]es / [A]lways Allow / [D]eny
You have three choices:
| Option | What it does |
|---|---|
| Yes (press Y) | Allow this one action. Claude will ask again next time. |
| Always Allow (press A) | Allow this tool for the rest of the session. Claude will not ask again for this specific tool. |
| Deny (press D) | Block the action. Claude will not perform it. |
If you are ever unsure about what Claude wants to do, choose Deny. Nothing bad happens — Claude will simply not perform that action, and you can always try again or ask Claude to explain what it was trying to do.
Always Allow lasts for your current session only. When you start a new Claude Code session, permissions reset and Claude will ask again. This is a safety feature.
This permission system is a core safety feature. Claude is powerful, but it never acts behind your back. You stay in control at all times.
Tool Use Indicators
When Claude is working, you will see it reference various tools. These are the built-in capabilities Claude uses to interact with your computer. Here is what each one does:
| Tool | What it does |
|---|---|
| Read | Reads the contents of a file so Claude can see what is in it |
| Write | Creates a new file or completely replaces an existing file |
| Edit | Makes targeted changes to a specific part of an existing file |
| Bash | Runs a terminal command (like npm install, git status, etc.) |
| Glob | Searches for files by name pattern (e.g., find all .js files) |
| Grep | Searches inside files for specific text or patterns |
| WebSearch | Searches the web for information |
| WebFetch | Fetches the contents of a specific URL |
When Claude uses a tool, you will see an indicator in the output showing which tool it is using and what it is doing with it. For example:
[Read] package.json
[Bash] npm install express
[Write] src/index.js
This helps you follow along with what Claude is doing at each step.
The Input Area
The input area at the bottom of the screen is where you type your messages. Here are the key shortcuts:
| Shortcut | What it does |
|---|---|
| Enter | Send your message |
| Shift+Enter | Start a new line without sending (for multi-line messages) |
| Up Arrow | Recall your previous message (like command history) |
| Escape or Ctrl+C | Cancel the current operation or interrupt Claude |
Shift+Enter is especially useful when you want to write a longer, more detailed prompt that spans multiple lines. Just hold Shift while pressing Enter to add a line break, then press Enter alone when you are ready to send.
Useful Commands While Chatting
Claude Code has several built-in slash commands you can type at any time during a conversation. These are not messages to Claude — they are commands for the Claude Code application itself.
| Command | What it does |
|---|---|
/help | Shows a list of available commands and keyboard shortcuts |
/clear | Clears the current conversation and starts fresh |
/cost | Shows how much of your usage quota this session has consumed |
/exit | Exits Claude Code (same as Ctrl+C twice) |
/compact | Summarizes the conversation to save context space — useful in long sessions when Claude starts to "forget" earlier messages |
Get in the habit of running /cost occasionally, especially when you are new. It helps you understand how different types of requests affect your usage.
Understanding Claude's Output
Claude's responses can contain several types of content:
- Text responses — Plain explanations, answers, and suggestions. This is the most common type of output.
- Tool calls — Indicators showing Claude is using a tool (like Read, Write, or Bash). These appear as highlighted blocks.
- Code blocks — Formatted blocks of code, often with syntax highlighting. Claude uses these when showing you code snippets or explaining what code does.
- Errors — If something goes wrong (a command fails, a file is not found), Claude will show the error and usually explain what happened and what to do next.
Verbose vs. Compact Mode
By default, Claude Code shows a condensed view of tool calls. If you want to see the full details of everything Claude is doing — the exact file contents it read, the complete output of commands it ran — you can toggle verbose mode.
Press Ctrl+O to toggle verbose mode on and off.
| Mode | What you see |
|---|---|
| Compact (default) | Brief summaries of tool calls — enough to follow along |
| Verbose | Full tool call details — every line of output, every file read |
Verbose mode is useful when you want to understand exactly what Claude did, or when you are debugging an issue and need to see the full output of a command.
Try It Yourself
Work through this short exercise to get comfortable with the interface:
- Start Claude Code — open PowerShell, navigate to any folder, and type
claude - Ask Claude to create a file — type:
Create a file called test-interface.txt with the text "Hello from Claude" - Watch the permission prompt — Claude will ask to use the Write tool. Choose Yes
- Check your costs — type
/costto see session usage - Get help — type
/helpto see all available commands - Toggle verbose mode — press Ctrl+O, then ask Claude to read the file you just created. Notice how much more detail you see. Press Ctrl+O again to return to compact mode
- Exit — type
/exitor press Ctrl+C twice
If you made it through all seven steps, you now know everything you need to navigate the Claude Code interface confidently.
What You Learned
- Claude Code is a text-based interface — no buttons or menus, just typing
- Claude always asks permission before acting on your computer
- You can allow, always allow, or deny any action
- Tools like Read, Write, Edit, Bash, Glob, and Grep are how Claude interacts with your files and system
- Slash commands like
/help,/clear,/cost,/exit, and/compactcontrol the application - Ctrl+O toggles verbose mode for full tool call details
How was this lesson? Take 2 minutes to share your feedback — it helps us make the tutorials better for everyone.