🖼️ @phantasm0009/image-zen

Local-first image optimizer with AI-powered processing - no cloud dependencies

🚀 Features

• 🎯 Zero-Cloud AI Tools: Background removal, super-resolution, and smart compression using TensorFlow Lite
• ⚡ Local Processing: Complete offline processing with pre-trained models
• 🛠️ CLI + API Support: Use from command line or integrate into your apps
• 🔗 Extensible Pipeline: Chain multiple operations seamlessly
• 🏃‍♂️ High Performance: Squoosh compression + TensorFlow Lite acceleration
• 📦 All-in-One: Pre-trained models downloaded automatically

🎯 Core Capabilities

| Feature | Technology | Performance | |---------|------------|-------------| | Background Removal | TensorFlow Lite + RobustVideoMatting | ~800ms avg | | Super-resolution | TensorFlow Lite + EDSR models | ~1.2s avg | | Smart Compression | Squoosh + multiple codecs | ~300ms avg | | Format Conversion | Support for JPG, PNG, WebP, AVIF | ~200ms avg |

📦 Installation

# Install globally for CLI usage
npm install -g @phantasm0009/image-zen

# Or install locally for API usage
npm install @phantasm0009/image-zen

Note: On first install, TensorFlow Lite models (~17MB total) will be downloaded automatically for offline AI processing.

🚀 Quick Start

CLI Usage

# Remove background from image
image-zen remove-bg photo.jpg -o clean.png

# Upscale image 2x with AI
image-zen upscale small.jpg --scale 2 -o large.png

# Compress with Squoosh optimization
image-zen compress *.jpg --quality 80 --format webp

# Enhanced pipeline processing
image-zen enhance photo.jpg -t remove-bg,upscale,compress -s 2 -q 85

API Usage

import { ImageZen, compress, removeBackground } from '@phantasm0009/image-zen';

// Create instance
const zen = new ImageZen({ verbose: true });

// AI background removal with TensorFlow Lite
const bgRemoved = await zen.removeBackground('photo.jpg');

// AI upscaling with EDSR models
const upscaled = await zen.upscale('small.jpg', 2);

// Squoosh compression
const compressed = await zen.compress('large.jpg', {
  quality: 80,
  format: 'webp'
});

// Enhanced pipeline
await zen.enhance('input.jpg', {
  tasks: ['remove-bg', 'upscale', 'compress'],
  scale: 2,
  compression: { quality: 85, format: 'webp' },
  output: 'output.webp'
});

// Direct function usage
const result = await compress('image.jpg', { quality: 70 });

🛠️ CLI Reference

Commands

remove-bg <input>

AI-powered background removal using TensorFlow Lite

image-zen remove-bg photo.jpg -o clean.png -f png

Options:

• -o, --output <path> - Output path or directory
• -f, --format <format> - Output format (png, webp) [default: png]

upscale <input>

AI super-resolution using EDSR TensorFlow Lite models

image-zen upscale small.jpg -s 4 -o large.png

Options:

• -s, --scale <number> - Scale factor (2, 4) [default: 2]
• -o, --output <path> - Output path or directory
• -f, --format <format> - Output format [default: png]

compress <input>

Smart compression using Squoosh library

image-zen compress *.jpg -q 80 -f webp -o compressed/

Options:

• -q, --quality <number> - Compression quality (1-100) [default: 80]
• -f, --format <format> - Output format (webp, avif, mozjpeg, oxipng) [default: webp]
• -o, --output <path> - Output path or directory

enhance <input>

Apply multiple AI enhancements in pipeline

image-zen enhance photo.jpg -t remove-bg,upscale,compress -s 2 -q 85

Options:

• -t, --tasks <tasks> - Comma-separated tasks [default: compress]
• -s, --scale <number> - Scale factor for upscaling [default: 2]
• -q, --quality <number> - Compression quality [default: 80]
• -f, --format <format> - Output format [default: webp]
• -o, --output <path> - Output path or directory

info

Show package information and model status

📚 API Reference

Class: ImageZen

Constructor

const zen = new ImageZen(options);

Options:

• verbose (boolean) - Enable verbose logging

Methods

removeBackground(input, options)

AI background removal using TensorFlow Lite

• input: string | Buffer - Image path or buffer
• options: Object - Processing options
• Returns: Promise<Buffer> - Processed image

upscale(input, scale, options)

AI upscaling using EDSR models

• input: string | Buffer - Image path or buffer
• scale: number - Scale factor (2 or 4)
• options: Object - Processing options
• Returns: Promise<Buffer> - Upscaled image

compress(input, options)

Smart compression using Squoosh

• input: string | Buffer - Image path or buffer
• options: Object - Compression options
  • quality: number (1-100) - Compression quality
  • format: string - Output format (webp, avif, mozjpeg, oxipng)
  • progressive: boolean - Progressive encoding
  • optimizeForSize: boolean - Size optimization
  • effort: number (0-6) - Compression effort
• Returns: Promise<Buffer> - Compressed image

enhance(input, config)

Enhanced processing pipeline

• input: string | Buffer - Image path or buffer
• config: Object - Processing configuration
  • tasks: Array - Processing tasks
  • output: string - Output file path
  • scale: number - Upscaling factor
  • compression: Object - Compression options
• Returns: Promise<Buffer | string> - Result or output path

🎨 Advanced Examples

AI-Powered Background Removal

const zen = new ImageZen({ verbose: true });

// High-quality background removal
const result = await zen.removeBackground('portrait.jpg', {
  format: 'png' // Preserve transparency
});

await fs.writeFile('portrait_clean.png', result);

Super-Resolution Upscaling

// 4x upscaling with AI enhancement
const upscaled = await zen.upscale('low_res.jpg', 4, {
  algorithm: 'lanczos3', // Fallback if TFLite model unavailable
  sharpen: true,
  enhanceDetails: true
});

Squoosh-Powered Compression

// Multi-format optimization
const formats = ['webp', 'avif', 'mozjpeg'];

for (const format of formats) {
  const compressed = await zen.compress('photo.jpg', {
    format,
    quality: await zen.getOptimalQuality('photo.jpg', format),
    optimizeForSize: true
  });
  
  await fs.writeFile(`photo.${format}`, compressed);
}

Batch Processing with Progress

const files = glob.sync('./photos/*.jpg');

const results = await zen.compressBatch(files, {
  format: 'webp',
  quality: 85
}, (current, total, result) => {
  console.log(`Progress: ${current}/${total} - ${result.file}`);
  if (result.success) {
    console.log(`  Saved ${result.compressionRatio}% space`);
  }
});

🔧 Technology Stack

AI Models (TensorFlow Lite)

• Background Removal: RobustVideoMatting (14.2MB)
• 2x Upscaling: EDSR 2x model (1.5MB)
• 4x Upscaling: EDSR 4x model (1.5MB)

Compression Engine

• Squoosh Library: Google's advanced image compression
• Supported Codecs: WebP, AVIF, MozJPEG, OxiPNG
• Optimization: Automatic quality selection and size optimization

Performance Features

• Tile-based Processing: Handle large images efficiently
• Memory Management: Automatic cleanup and optimization
• Batch Processing: Parallel processing with progress tracking
• Fallback Methods: Graceful degradation when models unavailable

📊 Performance Benchmarks

| Operation | TensorFlow Lite | Fallback | Memory Usage | |-----------|----------------|----------|--------------| | Background Removal | 800ms | 300ms | ~250MB | | 2x Upscaling | 1.2s | 400ms | ~300MB | | 4x Upscaling | 3.5s | 1.1s | ~600MB | | Squoosh Compression | 300ms | - | ~150MB |

Benchmarks on M1 MacBook Pro with 16GB RAM

🔧 Supported Formats

Input Formats

• JPEG (.jpg, .jpeg)
• PNG (.png)
• WebP (.webp)
• TIFF (.tiff)
• BMP (.bmp)

Output Formats

• WebP - Best overall compression
• AVIF - Next-gen format, excellent compression
• MozJPEG - Optimized JPEG with better compression
• OxiPNG - Optimized PNG with lossless compression

⚙️ Requirements

• Node.js: 16.0.0 or higher
• Memory: 2GB RAM minimum (4GB recommended for 4x upscaling)
• Storage: 150MB for package, models, and dependencies
• Internet: Required only for initial model download

🚀 Model Management

Models are automatically downloaded on first install:

# Check model status
image-zen info

# Manually trigger model download
npm run postinstall

# Models are stored in: node_modules/@phantasm0009/image-zen/models/

🤝 Contributing

We welcome contributions! Please see our Contributing Guide for details.

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests
5. Submit a pull request

📄 License

MIT License - see the LICENSE file for details.

💖 Support

If you find this package useful, consider:

• ⭐ Starring the repository
• 🐛 Reporting bugs
• 💡 Suggesting features
• 💝 Sponsoring development

🔗 Links

• NPM Package
• GitHub Repository
• Documentation
• Issue Tracker

Made with ❤️ by phantasm0009 | Local-first AI for everyone