Command-line interface reference for llm.do - Unified gateway for large language models (Large Language Models (LLMs))

Llm CLI

Unified gateway for large language models (Large Language Models (LLMs))

Installation

# Install globally
npm install -g do

# Or use via npx
npx do --version

Quick Start

# Basic usage
do integrations llm call /data

Commands

Main Command

do integrations llm call /data

Description: Unified gateway for large language models (Large Language Models (LLMs))

Usage:

do llm [options]

Options

Global Options

--help, -h - Show help
--version, -v - Show version
--verbose - Verbose output
--quiet, -q - Quiet mode
--json - JSON output
--config <file> - Config file path

Command-Specific Options

do llm \
  --option1=value1 \
  --option2=value2 \
  --verbose

Usage Examples

Basic Example

# Simple usage
do integrations llm call /data

With Options

# With configuration
do integrations llm call /data --verbose --json

Interactive Mode

# Interactive prompts
do llm --interactive

Batch Operations

# Process multiple items
do llm --batch items.json

Pipeline Integration

# Use with pipes
cat data.json | do llm --stdin | jq '.'

Configuration

Config File

Create .doconfig in your project:

{
  "llm": {
    "option1": "value1",
    "option2": "value2"
  }
}

Environment Variables

export DO_LLM_OPTION1=value1
export DO_LLM_OPTION2=value2

do integrations llm call /data

Output Formats

Default Output

do integrations llm call /data
# ✓ Operation completed
# Result: ...

JSON Output

do integrations llm call /data --json
# {
#   "success": true,
#   "result": { ... }
# }

Verbose Output

do integrations llm call /data --verbose
# → Starting operation...
# → Processing...
# → Completed in 1.2s

Scripting

Bash Scripts

#!/bin/bash

# Check if successful
if do integrations llm call /data; then
  echo "Success!"
else
  echo "Failed!"
  exit 1
fi

CI/CD Integration

# GitHub Actions
- name: Run Llm
  run: |
    do integrations llm call /data \
      --config prod.json \
      --quiet

Error Handling

Exit Codes

0 - Success
1 - General error
2 - Invalid arguments
3 - Timeout
4 - Not found

Error Messages

# Capture errors
do integrations llm call /data 2>errors.log

# Handle errors
if ! do integrations llm call /data; then
  echo "Error occurred, check errors.log"
  exit 1
fi

Debugging

# Debug mode
DEBUG=* do integrations llm call /data

# Dry run
do integrations llm call /data --dry-run

# Verbose + timing
time do integrations llm call /data --verbose

Best Practices

Use Config Files - Store configuration in .doconfig
Environment Variables - Use for sensitive data
Error Handling - Check exit codes in scripts
JSON Output - Use --json for parsing results
Verbose Mode - Use --verbose for debugging

Llm CLI

On this page