Humanizer — transparent AI editing + detector

This page shows a safe, transparent example: it edits (humanizes) text using an LLM and optionally checks the result with a detector API you provide. It will not attempt to hide that AI was used; instead it shows detector scores and keeps attribution visible.

Input text

Tone:

Length:

Humanized output — not yet generated

(no output yet)

Detector result

No detector run yet. To use detection, provide a detector backend endpoint (see instructions below).

Actions / transparency

Setup & safety notes

THIS TOOL IS INTENDED FOR TRANSPARENCY: it must not be used to deceive people about the origin of content. The UI shows when AI was used and shows detector results.
To make this page work you need a small backend (example below) that keeps your API keys secret. Do NOT put secret API keys in browser JavaScript.
The frontend calls two endpoints:
- /api/humanize — accepts POST JSON {text, tone, length} and returns {humanized}.
- /api/detect — accepts POST JSON {text} and returns {score:0-1, details} where higher score means "more likely AI". You can plug any detector you trust or host your own model.

Example backend (Node/Express) — create a file `server.js`

// Minimal example (do NOT expose your keys).
// Run: npm init -y && npm i express node-fetch

const express = require('express');
const fetch = require('node-fetch');
const app = express();
app.use(express.json());

// Read keys from env
const OPENAI_API_KEY = process.env.OPENAI_API_KEY; // for humanizing via ChatGPT
const DETECTOR_API_URL = process.env.DETECTOR_API_URL; // optional: your detector endpoint

app.post('/api/humanize', async (req, res) => {
  const { text = '', tone = 'conversational', length = 'same' } = req.body;
  if (!OPENAI_API_KEY) return res.status(500).json({ error: 'OpenAI API key not configured' });

  // Example using OpenAI Chat Completions (replace with your preferred call)
  const system = `You are a helpful editor. Rephrase the user's text to sound human, natural, and ${tone}. Keep meaning intact. If length is 'shorter' or 'longer', adjust accordingly.`;
  const userPrompt = `Edit this text for naturalness and human tone. Preserve meaning.\n\nText:\n"""${text}"""\n\nReturn only the edited text.`;

  const resp = await fetch('https://api.openai.com/v1/chat/completions', {
    method: 'POST',
    headers: { 'Content-Type': 'application/json', 'Authorization': `Bearer ${OPENAI_API_KEY}` },
    body: JSON.stringify({ model: 'gpt-4o-mini', messages: [{ role:'system', content: system }, { role:'user', content: userPrompt }], max_tokens: 800 })
  });
  const data = await resp.json();
  const humanized = data?.choices?.[0]?.message?.content ?? '';
  res.json({ humanized });
});

app.post('/api/detect', async (req, res) => {
  const { text = '' } = req.body;
  if (!DETECTOR_API_URL) return res.status(500).json({ error: 'No detector configured. Set DETECTOR_API_URL.' });
  // Proxy to your detector of choice. Detector must accept {text} and return {score}
  const resp = await fetch(DETECTOR_API_URL, { method:'POST', headers:{'Content-Type':'application/json'}, body:JSON.stringify({ text }) });
  const result = await resp.json();
  res.json(result);
});

app.listen(3000, () => console.log('Server on http://localhost:3000'));

// NOTES:
// - Do NOT use client-side API keys. Keep keys on server.
// - This example is intentionally generic about detectors: do not use it to avoid transparency.