Skip to content

CLI Options

Complete reference for command-line flags.

Usage

dial [options] [prompt]

General Options

Flag Short Description Default
--version -V Show version number -
--help -h Show help -
--debug -d Enable debug mode false
--verbose -v Verbose output false

Execution Options

Flag Short Description Default
--mode <mode> -m Execution mode (ask, quick, review, safe) auto
--yolo -y Skip all confirmations false
--sandbox -s Enable sandboxed execution false
--prompt <text> -p Run single prompt and exit -

Mode Values

Value Description
auto Automatic selection (default)
ask Read-only queries
quick Direct execution
review Light review cycle
safe Full review pipeline

Model Options

Flag Description Default
--model <name> Model to use From settings
--token-limit <n> Session token limit 32000

Authentication Options

Flag Description Default
--auth <type> Auth type (qwen-oauth, openai-compatible) qwen-oauth
--api-key <key> API key From env
--base-url <url> API base URL From env

Output Options

Flag Description Default
--json Output in JSON format false
--no-color Disable colored output false
--quiet Minimal output false

Session Options

Flag Description Default
--resume <id> Resume saved session -
--no-history Don't save to history false
--checkpoint Enable checkpointing true

Vision Options

Flag Description Default
--vlm-switch-mode <mode> Vision model switch behavior From settings

VLM Switch Values

Value Description
once Switch for one query
session Switch for session
persist Stay with vision model

IDE Options

Flag Description Default
--ide Enable IDE mode false

Examples

Basic Usage

dial                          # Interactive mode
dial "explain this code"      # Single prompt
dial -p "fix the bug"         # Prompt flag

Mode Selection

dial --mode safe              # Force safe mode
dial -m quick                 # Force quick mode
dial --yolo                   # Skip confirmations

Model Configuration

dial --model gpt-4            # Use specific model
dial --token-limit 16000      # Limit tokens

Output Control

dial --json                   # JSON output
dial --no-color               # No colors
dial --quiet                  # Minimal output

Headless Mode

dial -p "generate tests" --json > output.json
dial --prompt "fix bug" --quiet

Combining Options

# Safe mode with GPT-4
dial --mode safe --model gpt-4

# Headless JSON output
dial -p "analyze code" --json --quiet

# Debug with verbose output
dial --debug --verbose

Environment Precedence

CLI flags override environment variables, which override settings files:

  1. CLI flags (highest)
  2. Environment variables
  3. Project settings (.dial/settings.json)
  4. Global settings (~/.dial/settings.json)

Next Steps