🔒 CensorBot

A secure data sanitization tool for IT service companies that automatically detects and censors sensitive customer information using AI.

Overview

CensorBot is a Python application that helps protect customer privacy by automatically identifying and replacing sensitive information with placeholders. It uses a small, efficient LLM (like DeepSeek) to process text locally, ensuring that sensitive data never leaves your control before being sent to external AI services.

Features

• 🛡️ Automatic Detection - Identifies names, emails, phone numbers, addresses, SSNs, and more
• 🔄 Real-time Processing - Stream-based censoring for immediate feedback
• 🎯 High Accuracy - AI-powered detection understands context, not just patterns
• 💼 Enterprise Ready - Designed for IT service companies handling customer data
• 🌐 Web Interface - Clean, intuitive UI built with NiceGUI
• 📝 30+ Test Examples - Comprehensive test suite covering various scenarios

Quick Start

Prerequisites

• Python 3.8+
• uv package manager
• An OpenAI-compatible API endpoint (e.g., DeepSeek, local LLM)

Installation

1. Clone the repository:

git clone https://github.com/yourusername/CensorBot.git
cd CensorBot

2. Install dependencies:

uv sync

3. Configure environment variables:

cp .env.example .env
# Edit .env with your API credentials

4. Run the application:

uv run python src/main.py

5. Open your browser to http://localhost:8080

Configuration

Create a .env file with the following variables:

# LLM Backend Configuration
BACKEND_BASE_URL=https://api.deepseek.com  # Your LLM API endpoint
BACKEND_API_TOKEN=your-api-token-here       # API authentication token
BACKEND_MODEL=deepseek-chat                 # Model to use for censoring

Usage

1. Paste Text: Copy your text containing sensitive customer information into the input field
2. Process: Click "Censor Data" to automatically detect and replace sensitive information
3. Copy Result: Use the censored text safely with any external AI service

What Gets Censored

• Personal names
• Email addresses
• Phone numbers
• Physical addresses
• Social Security Numbers
• Credit card numbers
• Bank account numbers
• Driver's license numbers
• Passport numbers
• Medical record numbers
• IP addresses
• Usernames and passwords
• Company names (in customer context)
• Dates of birth

Project Structure

CensorBot/
├── src/
│   ├── main.py          # Main application with NiceGUI interface
│   ├── prompt.md        # System prompt for the censoring LLM
│   └── lib/
│       └── llm.py       # LLM integration module
├── examples/            # 30+ test cases with various sensitive data
│   ├── 01_customer_support.txt
│   ├── 02_medical_record.txt
│   └── ...
├── .env.example         # Environment variables template
├── pyproject.toml       # Project dependencies
└── CLAUDE.md           # AI assistant instructions

Development

Running Tests

Test the censoring with example files:

# The application loads a random example on startup
uv run python src/main.py

Adding Dependencies

uv add <package-name>

Project Commands

# Install dependencies
uv sync

# Run the application
uv run python src/main.py

# Format code (if configured)
uv run black src/

# Type checking (if configured)
uv run mypy src/

Security Considerations

• Local Processing: Use a local or self-hosted LLM for maximum security
• No Data Storage: CensorBot doesn't store any processed text
• API Security: Keep your API tokens secure and never commit them
• HTTPS Only: Use HTTPS for API communications
• Regular Updates: Keep dependencies updated for security patches

Use Cases

• IT Support Tickets: Sanitize customer tickets before using AI for solutions
• Documentation: Remove sensitive data from technical documentation
• Training Data: Prepare datasets for ML training without privacy concerns
• Compliance: Meet GDPR, HIPAA, and other privacy regulations
• Knowledge Base: Create sanitized versions of customer interactions

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add some AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Built with NiceGUI for the web interface
• Powered by uv for fast Python package management
• AI censoring via OpenAI-compatible APIs

Support

For issues, questions, or suggestions, please open an issue on GitHub.

---

⚠️ Important: This tool is designed to help protect privacy but should not be the only measure. Always review censored output and follow your organization's data protection policies.