ML Middleware

ML-powered request analysis middleware for Express and Next.js applications.

Installation

npm install rajpendkalkar123-ml-middleware

Quick Start

Express

import express from 'express';
import MLMiddleware from 'rajpendkalkar123-ml-middleware';

const app = express();
app.use(express.json());

const mlMiddleware = new MLMiddleware({
  mlApiUrl: 'https://your-ml-api.com/analyze',
  confidenceThreshold: 0.7,
  onSuspiciousRequest: async (req, res, classification) => {
    console.log('Blocked:', classification);
    res.status(403).json({ error: 'Request blocked' });
  }
});

app.use(mlMiddleware.middleware());

app.listen(3000);

Next.js

// middleware.js
import { MLMiddleware } from 'rajpendkalkar123-ml-middleware';

const mlMiddleware = new MLMiddleware({
  confidenceThreshold: 0.8
});

export async function middleware(request) {
  // Implement Next.js middleware logic
}

Configuration

• mlApiUrl: URL of your ML classification API
• confidenceThreshold: Minimum confidence to take action (0-1)
• analyzeBody: Analyze request body (default: true)
• analyzeQuery: Analyze query parameters (default: true)
• analyzeHeaders: Analyze headers (default: false)
• onSuspiciousRequest: Custom handler for suspicious requests

Features

• 🧠 ML-powered request analysis
• ⚡ Built-in caching for performance
• 🔄 Automatic retries on API failures
• 🛡️ Fail-open by default (traffic allowed if ML is unavailable)
• 📊 Confidence scoring
• 🎯 Customizable thresholds and handlers

License

MIT