Installing Claude Code (Claude in your terminal)
FEWD Studio works with any AI agent that writes to disk. Claude Code is the one we use ourselves, and it's the one most people ask about. This page explains what it actually is, why it behaves so differently from the AI you may have used inside a page builder, and how to get it installed.
If you just want the commands, skip to Installation.
What Claude Code is
Claude Code is a command-line tool from Anthropic. You open a terminal, navigate to a project folder, type claude, and you're in a session with an AI that can see and act on that project.
The important word is act. It isn't a chat window that hands you a snippet to copy. Within the directory you started it in, it can:
- read any file, and follow imports across the codebase
- create, edit and delete files
- run shell commands (your build, your dev server, a git diff, a search)
- read the output of those commands and respond to it
So a request like "the mobile nav is broken below 640px, find it and fix it" is something it can genuinely attempt. It greps for the relevant selectors, opens the files, makes the change, and can run your build to check nothing broke.
Why this is a different category to page-builder AI
Most designers and developers meet AI first inside a tool: a prompt box in WordPress, a generate button in a page builder, an assistant panel in a design app. Claude Code is not a bigger version of that. It's a different shape, and it's worth understanding why before you install it.
1. It has the whole project, not the current block
A prompt box inside a builder sees the element you're editing. It has no idea what your other components look like, what your token file contains, or what naming convention you've been using for the last six months.
An agent in your project root reads all of it. It can match your existing patterns because it can go and look at them. If you keep a CLAUDE.md or NOTES.md in the repo describing your conventions, it reads that on every session and works to it.
2. Its tools are primitives, not a feature list
This is the part people miss. Builder AI can only do what the vendor built an action for: generate a section, rewrite this copy, suggest a colour. If the thing you need isn't on that list, it can't be asked for.
Claude Code's tools are shell and filesystem. That means the scope isn't a menu, it's "anything you could do at a command line." Rename a class across forty files. Write a build script. Read a CSV and generate components from it. Nobody had to ship a feature for those.
3. The output is code you own
Files in your repo, in git, in a format any other tool can read. Not proprietary block data inside somebody's database. If you stop using the agent tomorrow, you keep everything it made.
A caveat
If you're not working in a codebase and don't want to be, a prompt box inside the tool you're already in is genuinely the better fit. Claude Code assumes you're comfortable in a terminal and that your work lives in a git repo.
How this fits with FEWD Studio
FEWD Studio and a terminal agent solve different halves of the same job, which is why they work well together rather than competing.
Claude Code is strong at reasoning across a whole codebase: refactors, wiring, scripts, anything that touches many files at once. It's weak at anything visual. It can't see your layout.
FEWD Studio is the opposite. The live node tree and round-trip visual editing give you direct manipulation of the thing you're actually looking at, with the result written straight back to disk as real code.
Because both write to plain files in the same repo, you can move between them freely. Restructure with the agent, then open FEWD and adjust the result by eye. There's no import step and no sync to manage. This is also why FEWD isn't tied to Claude Code specifically: Cursor, Gemini CLI or anything else that writes to disk will work the same way.
Before you start
You'll need:
- A supported OS. macOS 13+, Windows 10 (1809+), or Ubuntu 20.04+ / Debian 10+ / Alpine 3.19+.
- A terminal. Terminal or iTerm on macOS, PowerShell on Windows, whatever you already use on Linux.
- A paid Claude account. Claude Code requires a Pro, Max, Team, Enterprise or Anthropic Console account. The free Claude.ai plan does not include access.
You do not need Node.js. The installer below downloads a self-contained binary.
Installation
Step 1 — Run the installer
macOS, Linux, or WSL:
bash
curl -fsSL https://claude.ai/install.sh | bash
Windows PowerShell:
powershell
irm https://claude.ai/install.ps1 | iex
macOS via Homebrew
(alternative, if you prefer managing it through brew):
bash
brew install --cask claude-code
Note: the native installer keeps itself updated in the background. Homebrew installs do not, so you'd need to run brew upgrade claude-code yourself from time to time.
Step 2 — Fix your PATH if needed
The installer will usually finish with a "Setup notes" warning telling you that ~/.local/bin isn't on your PATH. If you skip this, claude will come back as command not found even though the install succeeded.
Run the line that matches your shell. If you're not sure which you're using, run echo $SHELL first.
zsh (the macOS default):
bash
echo 'export PATH="$HOME/.local/bin:$PATH"' >> ~/.zshrc source ~/.zshrc
bash:
bash
echo 'export PATH="$HOME/.local/bin:$PATH"' >> ~/.bash_profile source ~/.bash_profile
If you use both shells at different times, add it to both files. It's harmless to have it twice.
Step 3 — Verify the install
bash
claude --version
You should get a version number back, formatted like 2.1.252 (Claude Code).
If you want a fuller check of your setup, this prints diagnostics without starting a session:
bash
claude doctor
Step 4 — Start a session and log in
Navigate to a project folder and start it.
bash
cd ~/path/to/your-project claude
The first run opens your browser to authenticate. Log in with the Claude account from the prerequisites, and you're in.
This is easier to do in FEWD Studio. Once you’ve installed Claude, simply open a project with FEWD (or do a one-click start), open the Terminal at the top left, and you are already in your project folder, ready to go. Simply type:
claude
And Claude will fire right up!
Step 5 — Try it on something small
Once you're in a session, just type in plain English. A good first test on an existing project:
Read through this project and summarise the structure, the CSS conventions in use, and anything that looks inconsistent.
It costs you nothing but a minute, and it tells you immediately whether it has understood your codebase.
Useful commands
claude
Start an interactive session in the current directory
claude --version
Print the installed version
claude doctor
Diagnostics: install health, settings errors, update status
claude update
Apply an update immediately instead of waiting
/help
(inside a session) List available slash commands
Troubleshooting
claude: command not found right after a successful install. Almost always the PATH issue in Step 2. Run the appropriate export PATH line, then open a new terminal window and try again.
The install command errors out. If you're on Windows and see The token '&&' is not a valid statement separator, you're in PowerShell but ran the CMD command. If you see 'irm' is not recognized, you're in CMD but ran the PowerShell one. Your prompt shows PS C:\ in PowerShell.
It runs but won't authenticate. Confirm your Claude plan includes Claude Code. The free tier doesn't.
Anthropic's full setup and troubleshooting docs are at code.claude.com/docs/en/setup.
A note on cost
Claude Code uses a lot more tokens than a prompt box does. It reads whole files, holds large amounts of context, and loops through tool calls to check its own work. That's the mechanism by which it's useful.
What makes this workable is the pricing model rather than the efficiency. On a subscription plan, sustained agentic work is absorbed by a flat monthly fee rather than metered per generation. There are still session and weekly usage caps, and heavy use will find them, so it isn't unlimited. But it does mean the cost of asking it to do something ambitious is predictable.