CLI Reference

Complete command-line interface reference for Claude Code.

Installation

macOS / Linux / WSL

curl -fsSL https://claude.ai/install.sh | bash

Homebrew

brew install --cask claude-code

Windows PowerShell

irm https://claude.ai/install.ps1 | iex

Windows CMD

curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd && del install.cmd

NPM

npm install -g @anthropic-ai/claude-code

Basic Usage

Start Interactive Mode

claude

One-Time Task

claude "create a hello world function"

Query and Exit

claude -p "what files are in this directory"

Command-Line Options

Session Options

Flag	Description
-c, --continue	Continue most recent conversation
-r, --resume [id]	Resume specific session
--print	Print mode (non-interactive)

Output Options

Flag	Description
-p, --print	Print response and exit
--output-format <fmt>	Output format (text, json, stream-json)
--verbose	Verbose output
--debug	Debug mode with detailed logging

Model Options

Flag	Description
--model <model>	Specify model to use
--max-tokens <n>	Maximum output tokens

Permission Options

Flag	Description
--dangerously-skip-permissions	Skip permission prompts
--permission-mode <mode>	Set permission mode

MCP Options

Flag	Description
--mcp-config <file>	MCP configuration file
--mcp-debug	Debug MCP connections

Other Options

Flag	Description
--help	Show help
--version	Show version
--cwd <dir>	Set working directory
--config <file>	Configuration file

Examples

Basic Interaction

# Start interactive session
claude

# Ask a question
claude "explain this code" < main.py

# Continue last conversation
claude -c

Automation

# One-shot command
claude "fix the typescript errors in src/"

# Pipe input
cat error.log | claude "explain this error"

# JSON output
claude --output-format json "list all functions"

Advanced

# Custom model
claude --model claude-opus-4-5 "review this code"

# Debug mode
claude --debug "why is this failing"

# Skip permissions (CI/CD)
claude --dangerously-skip-permissions "run tests"

Environment Variables

Authentication

Variable	Description
ANTHROPIC_API_KEY	API key for direct access
CLAUDE_CODE_USE_BEDROCK	Enable AWS Bedrock
CLAUDE_CODE_USE_VERTEX	Enable Google Vertex AI

AWS Bedrock

Variable	Description
AWS_REGION	AWS region
AWS_PROFILE	AWS profile name
AWS_ACCESS_KEY_ID	Access key
AWS_SECRET_ACCESS_KEY	Secret key

Google Vertex AI

Variable	Description
ANTHROPIC_VERTEX_PROJECT_ID	GCP project ID
CLOUD_ML_REGION	Vertex AI region
GOOGLE_APPLICATION_CREDENTIALS	Service account path

Model Configuration

Variable	Description
ANTHROPIC_MODEL	Primary model
ANTHROPIC_SMALL_FAST_MODEL	Small/fast model
CLAUDE_CODE_MAX_OUTPUT_TOKENS	Max output tokens
MAX_THINKING_TOKENS	Max thinking tokens
DISABLE_PROMPT_CACHING	Disable caching

MCP

Variable	Description
MAX_MCP_OUTPUT_TOKENS	MCP output limit

Configuration Files

User Settings

~/.claude/settings.json:

{
  "model": "claude-sonnet-4-5",
  "theme": "dark",
  "permissions": {
    "autoApprove": ["Read", "Glob", "Grep"]
  }
}

Project Settings

.claude/settings.json:

{
  "model": "claude-opus-4-5",
  "allowedTools": ["Bash", "Edit", "Write"]
}

MCP Configuration

.mcp.json:

{
  "mcpServers": {
    "sqlite": {
      "command": "npx",
      "args": ["@anthropic/mcp-sqlite", "./db.sqlite"]
    }
  }
}

Exit Codes

Code	Meaning
0	Success
1	General error
2	Invalid arguments

Next Steps

• Slash Commands - Interactive commands
• Hooks - Automation
• MCP Guide - Tool integration