Getting Started

Quick Start

Build your first AI-powered application with Sigao AI DevKit in minutes.

Quick Start

Build your first AI-powered application with Sigao AI DevKit in minutes.

Prerequisites

Before starting, ensure you have:

  • ✅ Completed the Installation Guide
  • ✅ Node.js 18+ installed
  • ✅ Sigao CLI installed (npm install -g @sigaostudios/sigao-cli)
  • ✅ Claude CLI configured (installed via sigao setup)

🚀 Your First Workpackage

The fastest way to start is using our AI-powered workpackage system.

1. Create a Simple Requirements Document

Create a file called todo-app.md:

# Todo App Requirements

## Overview
Build a simple command-line todo application with the following features:

## Features
- Add new tasks
- List all tasks
- Mark tasks as complete
- Delete tasks
- Save tasks to a JSON file

## Technical Requirements
- Use Node.js
- Include unit tests
- Add a README with usage instructions

2. Generate a Workpackage

Convert your requirements into an executable workpackage:

sigao generate todo-app.md todo-workpackage.json

This creates a structured JSON file with:

  • 📋 Task breakdown
  • 🧪 Test definitions
  • 🔗 Dependencies
  • 📝 Context for Claude

3. Execute the Workpackage

Run the workpackage to build your application:

sigao work todo-workpackage.json --output-dir ./my-todo-app

Watch as Claude:

  • Creates the project structure
  • Implements each feature
  • Writes tests
  • Generates documentation

4. Verify Your Application

cd my-todo-app
npm install
npm test
node index.js --help

🎯 Quick Examples

Web API Project

# Create requirements
cat > api-requirements.md << EOF
# REST API for User Management
- CRUD operations for users
- JWT authentication
- Express.js framework
- MongoDB integration
- Swagger documentation
EOF

# Generate and execute
sigao generate api-requirements.md api-workpackage.json
sigao work api-workpackage.json --output-dir ./user-api

React Component Library

# Using a pre-built example
curl -O https://raw.githubusercontent.com/sigaostudios/sigao-ai-devkit/main/examples/workpackages/react-components.json
sigao work react-components.json --output-dir ./my-components

🛠️ Manual Setup (Traditional Approach)

If you prefer setting up projects manually:

1. Create Project Structure

mkdir my-ai-app && cd my-ai-app
npm init -y

2. Install Dependencies

npm install @anthropic-ai/sdk dotenv
npm install -D @types/node typescript jest

3. Create Basic AI Integration

Create index.js:

import Anthropic from '@anthropic-ai/sdk';
import dotenv from 'dotenv';

dotenv.config();

const anthropic = new Anthropic({
  apiKey: process.env.ANTHROPIC_API_KEY,
});

async function askClaude(prompt) {
  const message = await anthropic.messages.create({
    model: 'claude-3-opus-20240229',
    max_tokens: 1000,
    messages: [{ role: 'user', content: prompt }],
  });
  
  return message.content[0].text;
}

// Example usage
askClaude('Write a haiku about coding').then(console.log);

4. Set Up Environment

Create .env:

ANTHROPIC_API_KEY=your-api-key-here

5. Run Your Application

node index.js

📚 Sigao CLI Commands

Help System

# Show all available help topics
sigao help

# Get help for specific tools
sigao help git
sigao help work
sigao help docker

# Quick command reference
sigao cheat

Component Management

# List all available components
sigao --list

# Install specific components
sigao -c node,docker,python

# Dry run to preview changes
sigao --dry-run

Workpackage Commands

# Generate from PRD
sigao generate <input.md> <output.json>

# Execute workpackage
sigao work <package.json> [options]
  --model <model>      # Use specific Claude model
  --output-dir <path>  # Output directory
  --dry-run           # Preview without executing
  --verbose           # Show detailed logs

🔥 Pro Tips

1. Use the Right Model

# For complex tasks
sigao work package.json --model opus

# For faster iteration
sigao work package.json --model sonnet

2. Leverage Templates

Browse our example workpackages for common patterns.

3. Debug with Dry Run

# Preview what will happen
sigao work package.json --dry-run

# See detailed execution plan
sigao work package.json --verbose --dry-run

4. Customize Context

Add a CLAUDE.md file to your project for custom instructions that Claude will follow.

🎓 Next Steps

🆘 Getting Help

Navigation

Installation Guide

This guide walks you through setting up the Sigao AI DevKit and your development environment.

Core Concepts

Understanding the fundamental concepts behind AI-first development with Sigao AI DevKit.