Claude Code Complete Installation Guide: From Node.js Setup to MCP Server Integration

Claude Code Complete Installation Guide: First Run to Full Configuration

Claude Code is Anthropic’s agentic AI coding tool that operates directly in your terminal. This guide walks you through every step — from installing Node.js to configuring CLAUDE.md project instructions, connecting MCP servers, and setting up permission modes — so you can go from zero to a fully productive environment in one session.

Step 1: Install Node.js (v18+)

Claude Code requires Node.js version 18 or later. Verify your current version or install it fresh. # Check current version node —version

Or download directly from https://nodejs.org

Confirm the installation succeeded:

node —version

npm —version

Step 2: Install Claude Code Globally

Install Claude Code as a global npm package: npm install -g @anthropic-ai/claude-code

After installation, verify it works: claude --version

On first launch, Claude Code will prompt you to authenticate with your Anthropic account or API key. # Launch Claude Code claude

Or set your API key as an environment variable

export ANTHROPIC_API_KEY=YOUR_API_KEY

Step 3: Configure CLAUDE.md Project Instructions

CLAUDE.md files are persistent instruction files that Claude Code reads automatically on every session. They control how the AI behaves within your project.

File Hierarchy

• Dev server: npm run dev
• Run tests: npm run test
• Lint: npm run lint
• Build: npm run build EOF

Step 4: Set Up MCP Server Integration

Model Context Protocol (MCP) servers extend Claude Code's capabilities by connecting it to external tools and data sources. Configuration lives in JSON settings files.

MCP Configuration Locations

• Project-level: .claude/mcp.json in your project root- User-level: ~/.claude/mcp.json for global MCP servers

Example: Adding a Filesystem and Database MCP Server

# Create the project MCP config directory mkdir -p .claude

cat > .claude/mcp.json << ‘EOF’ { “mcpServers”: { “filesystem”: { “command”: “npx”, “args”: [“-y”, “@anthropic-ai/mcp-server-filesystem”, “/path/to/allowed/dir”] }, “postgres”: { “command”: “npx”, “args”: [“-y”, “@anthropic-ai/mcp-server-postgres”], “env”: { “DATABASE_URL”: “postgresql://user:password@localhost:5432/mydb” } }, “github”: { “command”: “npx”, “args”: [“-y”, “@anthropic-ai/mcp-server-github”], “env”: { “GITHUB_TOKEN”: “YOUR_API_KEY” } } } } EOF

After saving the config, restart Claude Code. You can verify connected servers by running: # Inside Claude Code session, type: /mcp

Step 5: Configure Permission Modes

Claude Code has three permission modes that control how much autonomy the AI has when executing tools:

claude config set autoApprove “Bash(npm run test),Read,Glob,Grep”

You can also configure permissions per-project in .claude/settings.json: { “autoApprove”: [ “Read”, “Glob”, “Grep”, “Bash(npm run test)”, “Bash(npm run lint)” ] }

Step 6: Your First Productive Session

With everything configured, start Claude Code in your project directory: cd /your/project claude

Try these workflows to validate your setup: - **Explore the codebase:** Ask Claude to summarize the project structure- **Run tests:** Ask Claude to run and analyze your test suite- **Fix a bug:** Point Claude at an issue and let it diagnose, fix, and verify- **Add a feature:** Describe what you need and let Claude implement it across files# Example: start a session with an initial prompt claude "Review the authentication module and identify any security issues"

Resume a previous conversation

claude —continue

Pro Tips for Power Users

• Use slash commands: Type /help inside a session to see all commands. /compact compresses conversation context when you are running long.- Pipe input: Feed files or command output directly — cat error.log | claude “What caused this crash?”- Multi-turn memory: Claude Code remembers context within a session. Chain requests without repeating background.- Custom slash commands: Create .claude/commands/ directory with markdown files for reusable prompt templates.- Headless mode for CI/CD: Use claude -p “Run tests and report” —output-format json in pipelines for structured output.- Use the Explore subagent: For complex codebase research, Claude Code dispatches a dedicated exploration agent that searches broadly without consuming your main conversation context.

Troubleshooting Common Errors

Can I use Claude Code without an Anthropic API key?

Yes. Claude Code supports authentication through your Anthropic account via an interactive login flow on first launch. You can also use it through Claude Pro, Team, or Enterprise subscriptions. The API key method is an alternative for users who prefer environment-variable-based auth or need headless/CI usage.

How do I add MCP servers without restarting Claude Code?

Currently, MCP server configuration changes require restarting your Claude Code session. After editing .claude/mcp.json, exit and relaunch Claude Code. You can verify your servers loaded correctly by running the /mcp slash command inside the session.

What is the difference between CLAUDE.md and .claude/settings.json?

CLAUDE.md contains natural-language instructions that guide the AI’s behavior — coding standards, project context, and workflow rules. The .claude/settings.json file is a structured configuration file that controls Claude Code’s operational settings like permission auto-approvals and tool configurations. Use CLAUDE.md for “how to code in this project” and settings.json for “how Claude Code should operate.”