LLM Data Generator

Universal data generator using Large Language Models. Generate structured JSON data with schemas or any text format (Markdown, CSV, emails, etc.) with LLM assistance.

Features

🎯 Schema-driven JSON generation

// Input schema
{
  "name": { "type": "string" },
  "price": { "type": "number", "minimum": 10, "maximum": 1000 }
}

// Generated output
{
  "name": "Wireless Gaming Mouse Pro X",
  "price": 89.99
}

📄 Text format generation

# Generate blog posts, emails, CSV data, SQL scripts, etc.
llmdatagen --config blog-posts.json

# Output: article_1.md
# How AI is Transforming Healthcare in 2025
# 
# The healthcare industry is experiencing a revolutionary...

🤖 Interactive CLI

$ llmdatagen

╔════════════════════════════════════════════╗
║       LLM Data Generator v1.0.8            ║
╚════════════════════════════════════════════╝

👤 Account Info:
  balance: $44.58 of $50.00 (10.8% used)
  
? Select LLM model:
  🤖 Auto Router (selects best available model)
❯ 🆓 Free    Meta: Llama 3.2 3B Instruct
  💰 $0.15/M OpenAI: GPT-4o mini
  
? How many items to generate? 20

📊 Generating 20 items...
  ✓ Generated item 1/20
  ✓ Generated item 2/20
  ...
  
✅ Generated 20 items to output/products.json
💰 Total cost: $0.0042

• Live model selection - Always up-to-date models from OpenRouter
• Progress tracking - Real-time generation status
• Cost estimation - Know costs before and after generation
• Validation - Automatic JSON Schema validation
• Extensible - Add custom configurations and templates

Installation

# Install globally
npm install -g @1337plus/llmdatagen

# Or use with npx without installation
npx @1337plus/llmdatagen

For alternative installation methods (from source, development setup, troubleshooting), see docs/installation.md.

Quick Start

Interactive Mode (Default)

# Start interactive wizard
llmdatagen

This will show a menu with options to:

• 🚀 Generate data
• ⚙️ Create/edit configuration
• 🌐 Open web interface
• 📊 Test API connection
• 📂 Open data folder

Web Interface

# Start web UI
llmdatagen web

# Or use alias
llmdatagen ui

The web interface runs on http://localhost:3000 and includes:

• Local Mode (localhost): Full features + admin panel for user management
• Cloud Mode: Multi-tenant with authentication (users created via admin panel)

Command Line

# Generate with specific config
llmdatagen generate --config myconfig.json

Note: The tool stores configurations and outputs in your Documents folder:

• Configurations: ~/Documents/llmdatagen/configs/
• Generated files: ~/Documents/llmdatagen/output/

For detailed usage instructions, command-line options, and advanced features, see docs/usage.md.

Output Formats

The generator supports two output formats:

JSON Format (default)

• ✅ Schema validation with JSON Schema
• ✅ Type-safe structured data
• ✅ Field-level generation control
• ✅ Perfect for APIs, databases, configurations

Text Format

• ✅ Any text-based format (Markdown, CSV, SQL, etc.)
• ✅ No schema constraints
• ✅ Template variables in prompts
• ✅ Perfect for content, documentation, scripts

Web Interface

Run the web UI for visual configuration and result viewing:

# Start the web server (default port 3000)
npm run web

# Start on custom port
PORT=8080 npm run web

# The server runs headless - open http://localhost:3000 in your browser

The web interface provides:

• 📁 Drag-and-drop configuration upload
• ✏️ Visual JSON editor with syntax validation
• 🎯 One-click data generation
• 📊 Result visualization
• 💰 Real-time account balance display
• 💬 Interactive chat with LLM models
• 👥 User management (admin panel - localhost only)
• 🔐 Multi-tenant support with isolated user data

Documentation

• Configuration Guide - Detailed guide on creating configurations
• Usage Guide - Command-line options and features
• Installation - Alternative installation methods
• Examples - Technical details for developers

Contributing

1. Fork the repository
2. Create your feature branch
3. Submit a pull request

License

MIT License - see LICENSE file for details