Moxra v1.0.0

🚀 Moxra v1.0.0 - Initial Release

Professional NSFW Content Detection Library for Images, GIFs, and Videos

---

📖 About This Release

Welcome to the first stable release of Moxra! A powerful, lightweight, and production-ready library for detecting inappropriate content in images, GIFs, and videos. Built with deep learning and ONNX Runtime, Moxra delivers 95%+ accuracy with lightning-fast inference speed.

---

✨ Features

Core Features

• ⚡ Blazing Fast - Detection in under 0.5 seconds
• 🎯 95%+ Accuracy - Trained on millions of labeled images
• 🖼️ Multi-format Support - Images, GIFs, and Videos
• 🕊️ Smart Veil Detection - Automatic hijab detection for Islamic content
• 🔒 Privacy First - 100% local processing, no data storage

Developer Experience

• 🌐 Public REST API - No authentication required
• 🎨 Modern Web UI - Interactive playground with dark/light mode
• 💻 Full CLI Tool - Command-line interface for batch processing
• 🐍 Python Library - Simple import and use
• ⚡ Async Support - Concurrent processing capabilities

Technical Excellence

• 🔌 Multi-model Support - Choose from 3 optimized models
• 💾 GPU Acceleration - CUDA and TensorRT support
• 📦 Automatic Model Download - No manual setup needed
• 🧹 Smart Memory Management - Automatic cache cleanup
• 🌍 Mirror Support - For users in China

---

📦 Model Files

These model files are required for detection. Please download them from the releases page:

Model	Description	Size	Download
moxra_model.onnx	Default MobileNet V2 Model	14 MB	Download
moxra_m2model.onnx	Optimized MobileNet V2 Model	14 MB	Download
moxra_i3model.onnx	High-Accuracy Inception V3 Model	92 MB	Download

---

📥 Installation

From PyPI (Recommended)

pip install moxra

With GPU Support

pip install moxra[gpu]

From Source

git clone https://github.com/moxra/moxra.git
cd moxra
pip install -e .

---

🚀 Quick Start

Python Library

from moxra import MoxraDetector

# Initialize detector
detector = MoxraDetector()

# Classify image with veil detection
result = detector.classify_with_veil("image.jpg")

# Check results
print(f"NSFW: {result['is_nsfw']}")
print(f"NSFW Score: {result['adjusted_nsfw_score']:.2%}")
print(f"Predictions: {result['predictions']}")

REST API Server

# Start the server
python -m moxra.api.app

# Or with uvicorn
uvicorn moxra.api.app:create_app --host 0.0.0.0 --port 8000

CLI Tool

# Basic usage
moxra -i image.jpg

# Advanced options
moxra -i video.mp4 --sample-rate 0.05 --max-frames 200 --format json --output result.json

---

📊 Detection Categories

Category	Description	Color
🟢 neutral	Safe and normal content	Green
🟡 sexy	Sexually suggestive content	Yellow
🔴 porn	Explicit pornographic content	Red
🟣 hentai	Anime explicit content	Purple
🔵 drawing	Artistic drawings and illustrations	Blue

---

🌐 API Endpoints

All endpoints are public - No authentication required!

Method	Endpoint	Description
POST	/api/v1/classify-img	Upload image (JPG, PNG, WEBP, BMP)
POST	/api/v1/classify-gif	Upload GIF animation
POST	/api/v1/classify-video	Upload video (MP4, AVI, MOV, MKV)
POST	/api/v1/classify-url	Analyze via URL
GET	/api/v1/health	Service health check
POST	/api/v1/cleanup	Memory cache cleanup

---

🖥️ CLI Options

Option	Description
-i, --input	Input file path (required)
-t, --type	Model type: d, m2, i3
-d, --device	Execution device: cpu, cuda, tensorrt
-s, --sample-rate	Video sampling rate (0 to 1)
-f, --max-frames	Maximum video frames
--format	Output format: json, pretty, simple
-o, --output	Save output to file
-v, --verbose	Verbose output

---

📁 Project Structure

moxra/
├── moxra/                    # Core package
│   ├── api/                  # REST API
│   ├── core/                 # Detection engine
│   ├── processors/           # Media processors
│   └── cli/                  # CLI interface
├── moxra_model/              # Model files
├── static/                   # Static assets
├── templates/                # HTML templates
├── run.py                    # Server launcher
├── setup.py                  # Package installer
└── README.md                 # Documentation

---

🔧 Requirements

• Python 3.8+
• ONNX Runtime 1.12+
• OpenCV 4.5+
• Pillow 9.0+
• NumPy 1.21+
• FastAPI 0.100+ (optional)
• Uvicorn 0.20+ (optional)

---

📊 Performance Benchmark

Format	Average Time	Max Size
Image (CPU)	0.3-0.5s	20 MB
Image (GPU)	0.1-0.2s	20 MB
GIF (100 frames)	2-3s	20 MB
Video (100 frames)	3-5s	100 MB
Video (GPU)	1-2s	100 MB

---

🕊️ Veil Detection

Moxra includes automatic veil (hijab) detection:

• 🔍 Automatic Detection - Identifies content with hijab
• 📊 Error Reduction - Up to 30% reduction in false positives
• 🎯 Confidence Score - Provides confidence level for veil detection
• 🎨 Multiple Methods - Uses color analysis for accurate detection

result = detector.classify_with_veil("image.jpg")
print(f"Veil detected: {result['veil']['has_veil']}")
print(f"Confidence: {result['veil']['confidence']}")

---

🤝 Contributing

We welcome contributions! Please follow these steps:

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

# Development setup
pip install -e ".[dev]"

# Run tests
pytest tests/

# Format code
black moxra/
isort moxra/

---

📄 License

MIT License - Free for personal and commercial use. See the LICENSE file for details.

---

👨‍💻 Developer

Abolfazl Zarei

---

⭐ Support

If you find this project useful, please consider:

• ⭐ Starring the repository
• 📢 Sharing with others
• 🐛 Reporting issues
• 🔧 Contributing to the code

---

📣 Social Media

Platform	Link
📱 Telegram Channel	@Ninja_Code
📱 Telegram ID	@Abolfazl_PGR
🟣 Rubika Channel	@Ninja_Code
🟣 Rubika ID	@NinjaCode
🐙 GitHub	AbolfazlZarei-dev
🌐 Website	abolfazlzarei.sbs
📧 Email	ninjacode.ir@gmail.com

---

Made with ❤️ by Abolfazl Zarei

Built with passion for the open-source community