AI Text Detector — FastAPI Backend

A simple REST API that detects whether a given text is AI-generated or Human-written.

---

Project Structure

project/
├── main.py           ← FastAPI app (this is the server)
├── model.joblib      ← your trained sklearn pipeline  ← YOU ADD THIS
├── requirements.txt
└── README.md

---

Setup

# 1. Install dependencies
pip install -r requirements.txt

# 2. Drop your trained model here
#    The file must be named exactly:  model.joblib

# 3. Start the server
uvicorn main:app --reload --port 8000

---

API Endpoints

Method	URL	Description
GET	/	API info
GET	/health	Health check
POST	/predict	Classify a text

POST /predict

Request body:

{
  "text": "The impact of artificial intelligence on modern society..."
}

Response:

{
  "label": 1,
  "category": "AI-generated",
  "confidence": 94.7
}

• label → 0 = Human, 1 = AI-generated
• category → human-readable string
• confidence → model confidence in % (0–100)

---

Interactive Docs

Once the server is running, open:

• Swagger UI → http://127.0.0.1:8000/docs
• ReDoc → http://127.0.0.1:8000/redoc

---

Frontend Integration

const API_URL = "http://127.0.0.1:8000";

async function detectText(text) {
  const res = await fetch(`${API_URL}/predict`, {
    method: "POST",
    headers: { "Content-Type": "application/json" },
    body: JSON.stringify({ text }),
  });
  return await res.json();
  // { label: 1, category: "AI-generated", confidence: 94.7 }
}