AI Humanizer Pro rewrites robotic, AI-generated text into natural writing that reads like a real person — and slips past AI detectors. It ships as a decoupled full-stack app:

• ⚡ FastAPI ASGI backend (API-only) — auth, API keys, usage analytics, OTP reset, admin console API, and the NLTK-powered humanizer.
• 🎨 Next.js 14 + Three.js frontend — an SEO-optimized, scroll-driven 3D landing page plus the humanizer tool, dashboard, docs, and a full admin console.

flowchart LR
    U([🧑 Browser]) -->|pages| N["▲ Next.js 14<br/>App Router"]
    N -->|backend proxy| F["⚡ FastAPI<br/>ASGI · Uvicorn"]
    N -.->|WebGL| T["🌌 Three.js<br/>scroll scene"]
    F --> DB[("🗄️ SQLite<br/>WAL")]
    F --> M["📧 SMTP<br/>OTP email"]
    F --> NL["📚 NLTK<br/>WordNet"]

    classDef front fill:#6366f1,stroke:#fff,color:#fff;
    classDef back fill:#009688,stroke:#fff,color:#fff;
    classDef store fill:#a855f7,stroke:#fff,color:#fff;
    class N,T front;
    class F,M,NL back;
    class DB store;

The Next.js dev server proxies every /backend/* request to FastAPI, keeping the session cookie same-origin (no CORS, no SameSite gymnastics).

ai-humanizer-pro/
├── app.py                 # FastAPI API (auth, keys, usage, admin, humanizer)
├── gunicorn.conf.py       # Uvicorn worker config for production
├── requirements.txt       # Python deps
├── humanizer.db           # SQLite (gitignored)
└── frontend/              # Next.js 14 + Three.js
    ├── next.config.js     # /backend/* → FastAPI proxy
    ├── tailwind.config.js
    └── src/
        ├── app/           # routes: /, /humanizer, /dashboard, /docs, /admin …
        ├── components/    # three/, landing/, admin/, auth/, ui/, layout/
        ├── context/       # AuthContext
        └── lib/           # api client, site config

Prerequisites: Python 3.11+ and Node.js 18+ (tested on 22).

python -m venv .venv
# Windows: .venv\Scripts\activate   |   macOS/Linux: source .venv/bin/activate
pip install -r requirements.txt
python app.py                       # → http://127.0.0.1:5001

cd frontend
npm install
npm run dev                         # → http://localhost:3000

Open http://localhost:3000 🎉

Backend

Frontend

POST /humanize — authenticate with a session cookie or an X-API-Key header.

{
  "text": "Your AI generated text here",
  "tone": "professional",
  "deep_mode": true
}

curl -X POST https://noai.devprithwiraj.in/humanize \
  -H "Content-Type: application/json" \
  -H "X-API-Key: hmnz_your_api_key_here" \
  -d '{"text":"...","tone":"professional","deep_mode":true}'

const res = await fetch("https://noai.devprithwiraj.in/humanize", {
  method: "POST",
  headers: { "Content-Type": "application/json", "X-API-Key": "hmnz_..." },
  body: JSON.stringify({ text: "...", tone: "professional", deep_mode: true }),
});
const data = await res.json();
console.log(data.humanized);

sequenceDiagram
    participant C as Client
    participant F as FastAPI
    participant N as NLTK
    C->>F: POST /humanize (X-API-Key)
    F->>F: verify key / session
    F->>N: rephrase + perplexity spike
    N-->>F: humanized text
    F->>F: diff metrics + log usage
    F-->>C: { humanized, ai_probability, diff_html, ... }

One service — website + API on a single URL (Docker):

• Render → New + → Blueprint (uses render.yaml + Dockerfile), or set the service's Runtime = Docker (./Dockerfile).
• The image builds the Next.js static export and serves it with FastAPI — no separate frontend host.
• Set ADMIN_EMAIL, ADMIN_PASSWORD, SENDER_EMAIL, SENDER_PASSWORD in the dashboard (SECRET_KEY is auto-generated). Health check: /healthz.

🔴 Note: SQLite on ephemeral hosts resets on redeploy — use managed Postgres or a persistent disk for durable data.

• FastAPI API-only backend
• Next.js + Three.js frontend
• Admin console in Next.js
• SEO (metadata, JSON-LD, sitemap)
• Migrate SQLite → Postgres
• Rate limiting & quotas
• Stripe billing tiers

Prithwiraj 🌐 noai.devprithwiraj.in

⭐ If this project helped you, give it a star! ⭐