OpenRouter AI Mock Generator

AI-powered mock data generator with advanced image processing capabilities. Generate realistic mock data from existing JSON examples using state-of-the-art language models, with optional AI-generated images.

✨ Features

🤖 AI-Powered Generation: Uses OpenRouter API with JSON mode for reliable output
📊 Smart Pattern Analysis: Automatically detects and analyzes JSON structures
⚡ Dynamic Batch Processing: Optimizes generation based on data complexity
🎨 Custom Preferences: Guide generation with custom instructions
🖼️ AI Image Processing: Replace placeholder images with contextual AI-generated images
☁️ Cloud Storage Integration: Upload images to AWS S3 or DigitalOcean Spaces
🧠 Smart Optimization: Intelligent duplication detection and image reuse (up to 30% savings)
🌍 Italian Locale Support: Default Italian names, addresses, and formatting
📈 Multiple Model Support: Works with OpenAI, Anthropic Claude, Google Gemini, and more

Installation

npm install @bramato/openrouter-mock-generator

🚀 Quick Start

1. Setup Configuration

Run the interactive configuration wizard:

npx ai-init

This will guide you through:

OpenRouter API Key: Required for text generation
Model Selection: Choose the best models for your needs
AI Image Processing (Optional):
- Hugging Face API key for AI image generation
- Cloud storage (AWS S3 or DigitalOcean Spaces) for image hosting

2. Generate Mock Data

# Basic usage (uses placeholder images)
npx ai-generate-mock products.json --count 50

# With AI image processing (if configured)
npx ai-generate-mock products.json --images --count 20

# Force disable AI images
npx ai-generate-mock products.json --no-images --count 50

📋 CLI Usage

Basic Commands

# Generate mock data with auto-detection
npx ai-generate-mock <input.json> --count <number>

# With custom preferences
npx ai-generate-mock products.json --preferences "Generate luxury Italian fashion items" --count 25

# Specify output file and array path
npx ai-generate-mock data.json --output generated_data.json --array-path "products" --count 100

# Analyze file structure first
npx ai-generate-mock input.json --analyze

AI Image Processing

# Enable AI image processing (replaces Picsum URLs with AI-generated images)
npx ai-generate-mock products.json --images --count 50

# Disable AI image processing (keeps placeholder images)
npx ai-generate-mock products.json --no-images --count 50

# Auto-detection (enabled if configured, disabled otherwise)
npx ai-generate-mock products.json --count 50

CLI Options

-c, --count <number>: Number of items to generate (default: 10)
-o, --output <file>: Output file path (default: input_mock.json)
-p, --array-path <path>: Specific array path to use (e.g. 'data.items')
--preferences, --pref <text>: Custom preferences for data generation
--images, --enable-images: Force enable AI image processing
--no-images, --disable-images: Force disable AI image processing
-a, --analyze: Analyze input file structure without generating
-h, --help: Show help message

💻 Programmatic Usage

Basic Usage

import { MockGeneratorService } from '@bramato/openrouter-mock-generator';

// Auto-detect image processing configuration
const generator = new MockGeneratorService();

const result = await generator.generateMockData({
  inputFile: 'products.json',
  outputFile: 'generated_products.json',
  count: 50,
  preferences: 'Generate Italian fashion brands with prices in euros'
});

if (result.success) {
  console.log(`Generated ${result.generatedCount} items`);
} else {
  console.error(`Error: ${result.error}`);
}

Advanced Configuration

import { MockGeneratorService } from '@bramato/openrouter-mock-generator';

// Explicitly control image processing
const generator = new MockGeneratorService(
  undefined, // Use default OpenRouter config
  true       // Force enable image processing
);

const result = await generator.generateMockData({
  inputFile: 'products.json',
  outputFile: 'generated_products.json',
  count: 50,
  preferences: 'Generate luxury Italian fashion items',
  enableImageProcessing: true // Enable for this specific request
});

Image Processing Only

import { PostProcessingOrchestrator } from '@bramato/openrouter-mock-generator';

// Process existing mock data with AI images
const processor = new PostProcessingOrchestrator('your-hf-api-key', {
  verbose: true,
  enableOptimization: true,
  uploadToCloud: true
});

const data = JSON.parse(fs.readFileSync('existing-mock-data.json', 'utf8'));
const result = await processor.processData(data);

if (result.success) {
  fs.writeFileSync('processed-data.json', JSON.stringify(result.processedData, null, 2));
}

🔧 Environment Variables

Required

OPENROUTER_API_KEY: Your OpenRouter API key

Model Configuration

OPENROUTER_DEFAULT_MODEL: Default model to use
OPENROUTER_MOCK_GENERATOR_MODEL: Specific model for mock generation

AI Image Processing (Optional)

HUGGINGFACE_API_KEY: Hugging Face API token for image generation
STORAGE_PROVIDER: aws or digitalocean

AWS S3 Configuration

AWS_ACCESS_KEY_ID: AWS access key
AWS_SECRET_ACCESS_KEY: AWS secret key
AWS_REGION: AWS region (e.g., us-east-1)
AWS_S3_BUCKET_NAME: S3 bucket name
AWS_S3_ENDPOINT: Custom S3 endpoint (optional)

DigitalOcean Spaces Configuration

DO_SPACES_ACCESS_KEY: Spaces access key
DO_SPACES_SECRET_KEY: Spaces secret key
DO_SPACES_REGION: Spaces region (e.g., fra1)
DO_SPACES_NAME: Space name
DO_API_TOKEN: DigitalOcean API token (optional)

Supported Models

The package automatically filters and uses only JSON-mode compatible models:

OpenAI: GPT-4.1 Nano, GPT-4o Mini, GPT-4 Turbo, GPT-3.5 Turbo
Anthropic: Claude 3.5 Sonnet, Claude 3 Opus, Claude 3 Haiku
Google: Gemini Pro, Gemini Flash
And many more via OpenRouter

📚 Examples

E-commerce Product Catalog

# Input: products.json with Picsum placeholder images
npx ai-generate-mock products.json --images --preferences "Generate luxury Italian fashion items from Milano" --count 25

# Output: Real product images uploaded to cloud storage
# - Product photos with Italian fashion context
# - Thumbnails and gallery images  
# - Banner images for promotional content

Restaurant Menu with Food Images

npx ai-generate-mock menu.json --images --preferences "Traditional Italian cuisine from different regions" --count 15

# Generates:
# - Professional food photography for dishes
# - Restaurant interior images
# - Chef portrait photos

User Profiles with Avatars

npx ai-generate-mock users.json --images --preferences "Italian professionals from major cities" --count 50

# Creates:
# - Professional headshot portraits
# - Contextual background images
# - Company logos for businesses

🎯 Use Cases

Development & Testing

Generate realistic test data for frontend applications
Create demo content for presentations and prototypes
Populate databases with meaningful sample data

Design & Prototyping

Generate product catalogs with real imagery
Create user interfaces with contextual content
Build complete application mockups

Marketing & Content

Generate sample content for marketing campaigns
Create diverse product showcases
Build realistic user testimonials and reviews

💡 Custom Preferences Examples

# E-commerce products
--preferences "Generate luxury Italian fashion items with designer brands and premium pricing"

# Restaurant menu
--preferences "Create traditional Italian dishes with regional specialties and wine pairings"

# User data  
--preferences "Generate Italian professionals from Milano, Roma, and Napoli with realistic demographics"

# Events
--preferences "Create cultural events and festivals typical of Italian regions with local traditions"

# Real estate
--preferences "Generate Italian properties in historic city centers with authentic architectural details"

Smart Features

Dynamic Batch Sizing

The generator analyzes data complexity and automatically adjusts batch sizes:

Simple items: Up to 10 per batch
Complex nested structures: 1-3 per batch
Prevents API timeouts and ensures quality

🎨 AI Image Processing

The system can automatically replace placeholder images (Picsum URLs) with contextually appropriate AI-generated images:

How It Works

Detection: Scans JSON data for Picsum URLs (https://picsum.photos/...)
Context Analysis: Analyzes surrounding data to understand image context
Smart Optimization: Groups similar images to reduce generation costs
AI Generation: Creates images using Hugging Face models (FLUX.1-dev, Qwen-Image)
Cloud Upload: Uploads to AWS S3 or DigitalOcean Spaces with CDN URLs
URL Replacement: Replaces original URLs with new cloud URLs

Supported Image Types

Product Images: Professional product photography
Avatars & Profiles: Portrait photography
Banners & Headers: Wide format promotional images
Thumbnails: Small preview images
Logos & Icons: Brand identity elements
Backgrounds: Texture and pattern images

Optimization Features

Smart Duplication Detection: Reuses similar images (up to 30% cost savings)
Contextual Descriptions: Generated in Italian with local context
Batch Processing: Concurrent generation with rate limiting
Fallback Support: Graceful degradation if generation fails

Italian Locale Support

Realistic Italian names and surnames
Italian addresses and cities
Euro currency formatting
Italian business names and contexts

🔧 Troubleshooting

Image Processing Issues

Problem: AI image processing not working

# Check configuration
npx ai-init

# Test with explicit flag
npx ai-generate-mock data.json --images --count 1

Problem: Images not uploading to cloud storage

# Verify storage configuration
cat .env | grep -E "(STORAGE_PROVIDER|AWS_|DO_)"

# Test cloud connectivity
npx spaces-manager test  # For DigitalOcean

Problem: Hugging Face API errors

# Verify API key permissions at https://huggingface.co/settings/tokens
# Ensure "Make calls to the serverless Inference API" is enabled

Generation Issues

Problem: JSON parsing errors

Check input file format with npx ai-generate-mock input.json --analyze
Ensure input contains valid JSON arrays

Problem: Rate limiting

Reduce batch size with smaller --count values
Check OpenRouter/Hugging Face usage limits

Configuration Issues

Problem: Environment variables not loading

# Check .env file format (no spaces around =)
OPENROUTER_API_KEY=your-key-here

# Not: OPENROUTER_API_KEY = your-key-here

❓ FAQ

Q: Do I need both OpenRouter and Hugging Face API keys? A: OpenRouter is required for text generation. Hugging Face is optional for AI image processing.

Q: Can I use AWS S3 and DigitalOcean Spaces simultaneously? A: No, choose one storage provider in the configuration.

Q: How much does image generation cost? A: Hugging Face Inference API has free tier with rate limits. Cloud storage costs depend on usage.

Q: Can I process existing mock data with AI images? A: Yes, use the PostProcessingOrchestrator programmatically or re-run generation with --images.

Q: What image formats are supported? A: The system generates PNG images and supports various dimensions based on context.

Q: How does the optimization work? A: The system detects similar images by content and context, reusing generated images to reduce costs.

🚀 Advanced Features

Microservices Architecture

The image processing system uses a modular microservices approach:

ImageUrlExtractor: Finds and categorizes image URLs
DescriptionGenerator: Creates contextual AI prompts
ImageProcessingAnalyzer: Optimizes generation strategy
UrlReplacer: Updates JSON with new image URLs
PostProcessingOrchestrator: Coordinates all services

Cloud Storage Features

CDN Integration: Automatic CDN URL generation
Public Access: Images configured for public viewing
Organized Structure: Date-based folder organization
Fallback Support: Graceful handling of upload failures

📊 Performance

Optimization: Up to 30% cost savings through smart image reuse
Concurrency: Configurable concurrent image generation
Batch Processing: Automatic batching based on data complexity
Rate Limiting: Built-in protection against API limits

License

MIT

Author

bramato

🌟 If this project helps you, please consider starring it on GitHub!

Name		Name	Last commit message	Last commit date
src		src
.gitignore		.gitignore
README.md		README.md
bun.lock		bun.lock
mock-generated-data.json		mock-generated-data.json
package-lock.json		package-lock.json
package.json		package.json
risultato.json		risultato.json
test-output-simple.json		test-output-simple.json
test-output.json		test-output.json
test-post-processed-output.json		test-post-processed-output.json
test-post-processing.js		test-post-processing.js
tsconfig.json		tsconfig.json

Latest commit

History

Folders and files

Repository files navigation