A security proxy for LLM APIs with real-time PII detection and redaction. Protect your organization from accidentally sending sensitive data to language models.

• 🔒 Real-time PII Detection: Automatically detects and redacts sensitive information
• 🚀 Easy Integration: Simply change your API endpoint URL
• 📊 Comprehensive Logging: Track PII incidents and API usage
• 🐳 Self-Hosting Ready: Deploy with Docker or npm
• ⚡ High Performance: Minimal latency overhead
• 🔧 Configurable: Customize detection rules and behavior

• Social Security Numbers (SSN)
• Credit Card Numbers
• Email Addresses
• Phone Numbers
• IP Addresses
• API Keys
• Custom Patterns (configurable)

npm install -g llm-firewall

llm-firewall start

Your proxy is now running on http://localhost:3000

Simply replace your LLM API endpoint:

// Before (direct to OpenAI)
const response = await fetch('https://api.openai.com/v1/chat/completions', {
  headers: { 'Authorization': 'Bearer YOUR_KEY' }
});

// After (through LLM Firewall)
const response = await fetch('http://localhost:3000/proxy/openai/v1/chat/completions', {
  headers: { 'Authorization': 'Bearer YOUR_KEY' }
});

const axios = require('axios');

const response = await axios.post('http://localhost:3000/proxy/openai/v1/chat/completions', {
  model: 'gpt-3.5-turbo',
  messages: [
    {
      role: 'user',
      content: 'Help me with user info: email john@example.com, SSN 123-45-6789'
    }
  ]
}, {
  headers: {
    'Authorization': 'Bearer YOUR_OPENAI_API_KEY',
    'Content-Type': 'application/json'
  }
});

The request sent to OpenAI will be:

Help me with user info: email [REDACTED_EMAIL], SSN [REDACTED_SSN]

const response = await axios.post('http://localhost:3000/proxy/anthropic/v1/messages', {
  model: 'claude-3-haiku-20240307',
  max_tokens: 1000,
  messages: [
    {
      role: 'user',
      content: 'Process this customer data: phone 555-123-4567'
    }
  ]
}, {
  headers: {
    'x-api-key': 'YOUR_ANTHROPIC_API_KEY',
    'Content-Type': 'application/json',
    'anthropic-version': '2023-06-01'
  }
});

llm-firewall init

Create a config.json file:

{
  "port": 3000,
  "pii": {
    "detectSSN": true,
    "detectCreditCards": true,
    "detectEmails": true,
    "detectPhoneNumbers": true,
    "detectIPAddresses": true,
    "detectAPIKeys": true,
    "customPatterns": [
      {
        "name": "Employee ID",
        "pattern": "EMP-\\d{6}",
        "replacement": "[REDACTED_EMPLOYEE_ID]"
      }
    ]
  },
  "logging": {
    "level": "info",
    "logPII": false,
    "logRequests": true
  }
}

export LLM_FIREWALL_PORT=3000
export LLM_FIREWALL_LOG_LEVEL=info
export LLM_FIREWALL_DETECT_SSN=true
export LLM_FIREWALL_DETECT_EMAILS=true

version: '3.8'

services:
  llm-firewall:
    image: llm-firewall:latest
    ports:
      - "3000:3000"
    environment:
      - LLM_FIREWALL_PORT=3000
      - LLM_FIREWALL_LOG_LEVEL=info
    volumes:
      - ./logs:/app/logs
      - ./config:/app/config
    restart: unless-stopped

docker-compose up -d

# Build
docker build -t llm-firewall .

# Run
docker run -p 3000:3000 -d llm-firewall

llm-firewall start [options]

Options:
  -p, --port <port>     Port to run the server on (default: 3000)
  -c, --config <path>   Path to config file

 node dist/cli.js start

llm-firewall init [options]

Options:
  -o, --output <path>   Output path for config file

llm-firewall test [options]

Options:
  -t, --text <text>     Text to test for PII
  -c, --config <path>   Path to config file

curl http://localhost:3000/health

curl http://localhost:3000/status

• No Data Persistence: PII is redacted in memory, not stored
• Configurable Logging: Choose what to log (PII incidents vs. content)
• Rate Limiting: Built-in protection against abuse
• HTTPS Support: TLS termination ready
• Audit Trails: Comprehensive logging for compliance

• Low Latency: ~10ms overhead for PII detection
• High Throughput: Supports concurrent requests
• Memory Efficient: Minimal memory footprint
• Scalable: Stateless design for horizontal scaling

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

MIT License - see LICENSE file for details

• Advanced NLP-based PII detection
• Web dashboard for monitoring
• Custom redaction rules
• Integration with identity providers
• Compliance reporting (GDPR, HIPAA)
• Multi-tenant support