ClueoBots

AI security for your app. One line of code.

Install

npm install @openclueo/clueobots

Setup

1. Get your API key: https://clueobots.com/api-keys

2. Add to .env:

CLUEOBOTS_API_KEY=sk_cb_your_key_here

3. Choose your integration style:

Simple Direct Scanning (Manual Control)

import { scan } from '@openclueo/clueobots'

// Scan any text content
const result = await scan("User input to check")
if (result.threatDetected) {
  console.log(`Threat found: ${result.primaryThreat}`)
}

Middleware Integration (Automatic Scanning)

Express.js:

import { expressMiddleware } from '@openclueo/clueobots'
import express from 'express'

const app = express()
app.use(expressMiddleware()) // Scans all JSON responses automatically

Fastify:

import { fastifyPlugin } from '@openclueo/clueobots'

app.register(fastifyPlugin())

Koa:

import { koaMiddleware } from '@openclueo/clueobots'

app.use(koaMiddleware())

Legacy Auto-Detection (Zero Config)

CommonJS:

require('@openclueo/clueobots')() // Auto-detects framework

ES Modules:

import clueobots from '@openclueo/clueobots'
clueobots() // Auto-detects framework

Next.js (Manual Output Scanning)

import { scanOutput } from '@openclueo/clueobots/nextjs'

export async function POST(req: Request) {
  const { message } = await req.json()

  // Generate AI response
  const aiResponse = await openai.chat.completions.create({
    messages: [{ role: 'user', content: message }]
  })

  const content = aiResponse.choices[0].message.content

  // Scan output before sending
  const scan = await scanOutput(content)
  if (!scan.allowed) {
    return Response.json({ error: 'Response blocked for security' }, { status: 403 })
  }

  return Response.json({ message: content })
}

Done. ClueoBots is now protecting your AI app from:

• 🛡️ Prompt injection attacks (that succeed)
• 🔒 Data exfiltration in AI responses
• ⚠️ Jailbreak attempts that bypass your defenses
• 🚨 Harmful content in AI outputs

Questions? [email protected]

Docs: https://clueoai.com/docs

License: MIT