Skip to main content

​
Introduction

Qwen Code is an AI-powered coding assistant that runs directly in your terminal. The CLI supports both interactive and headless modes, making it suitable for development workflows, automation, and CI/CD pipelines.

​
Basic Usage

​
Interactive Mode

Launch Qwen Code in interactive mode for a conversational coding experience: 
qwen
This opens an interactive terminal interface where you can chat with the AI, execute commands, and work on your code.

​
Headless Mode

Use the --prompt flag to run Qwen Code in non-interactive mode: 
qwen --prompt "Create a Python function to calculate fibonacci numbers"
You can also pipe input from stdin: 
echo "Explain this code" | qwen --prompt

​
Positional Arguments

Provide a prompt as positional arguments for one-shot execution: 
qwen "Write unit tests for my authentication module"
Use --prompt-interactive to execute a prompt and continue in interactive mode: 
qwen --prompt-interactive "Setup a React project with TypeScript"

​
Available Commands

Qwen Code provides slash commands that you can use within interactive mode or with the --prompt flag:

/auth

Configure authentication settings

/help

Display help information

/clear

Clear conversation history

/compress

Compress conversation context

/model

Switch the active model

/stats

View session statistics

/bug

Submit a bug report

​
Command-Line Options

Configure Qwen Code behavior with command-line flags:

--prompt

Execute a prompt in non-interactive mode

--model

Specify the AI model to use

--yolo

Auto-approve all actions

--plan

Plan-only mode without execution

--resume

Resume a previous session

--session-id

Specify a custom session ID

​
Output Formats

Qwen Code supports multiple output formats for different use cases:

​
Text Format (Default)

Human-readable output suitable for terminal usage: 
qwen --prompt "Hello"

​
JSON Format

Structured output for programmatic consumption: 
qwen --prompt "Analyze this code" --output-format json

​
Stream JSON Format

Real-time streaming output with JSON events: 
qwen --prompt "Generate a report" --output-format stream-json

​
Input Formats

Control how Qwen Code processes input:
  • text: Standard text input (default)
  • stream-json: JSON-formatted streaming input for programmatic control
qwen --input-format stream-json

​
Session Management

​
Starting a New Session

Each invocation creates a new session by default: 
qwen

​
Resuming Sessions

Resume the most recent session: 
qwen --continue
Resume a specific session by ID: 
qwen --resume <session-id>
Show session picker: 
qwen --resume

​
Custom Session ID

Specify a custom session ID: 
qwen --session-id my-custom-session

​
Approval Modes

Control how Qwen Code handles tool execution:
  • default: Prompt for approval before executing tools
  • yolo: Auto-approve all tool executions
  • plan: Generate plans without executing
  • auto-edit: Auto-approve edit operations only
qwen --approval-mode yolo

​
Examples

​
Code Generation

qwen --prompt "Create a REST API with Express.js"

​
Code Review

git diff | qwen --prompt "Review these changes"

​
Automated Testing

qwen --yolo --prompt "Add unit tests to all functions in src/"

​
CI/CD Integration

qwen --prompt "Run linter and fix all issues" --output-format json

​
Environment Variables

Configure Qwen Code using environment variables:
  • QWEN_CODE_NO_RELAUNCH: Disable automatic memory reconfiguration
  • NO_COLOR: Disable colored output
  • DEBUG: Enable debug logging
  • SANDBOX: Indicates running inside sandbox

​
Configuration Files

Qwen Code reads configuration from:
  • Global settings: ~/.qwen/settings.json
  • Project settings: .qwen/settings.json
  • Project config: .qwen.config.js
See the Configuration guide for details.

​
Getting Help

View all available options: 
qwen --help
Access help within interactive mode: 
/help
Submit bug reports: 
/bug "Description of the issue"

​
Next Steps

Authentication

Set up API keys and OAuth

Configuration

Customize Qwen Code settings

Features

Explore powerful features

IDE Integration

Use Qwen Code in your editor