Text Humanizer & AI Detection API

A FastAPI-based backend service for humanizing AI-generated text and detecting AI content.

• Analyzes text to detect AI-generated content
• Classifies each sentence as:
  • AI-generated
  • AI-generated & AI-refined
  • Human-written
  • Human-written & AI-refined
• Returns detailed percentages and classification results
• Uses roberta-base-openai-detector model

• Humanizes AI-generated text while preserving APA citations
• Expands contractions naturally
• Replaces words with contextual synonyms using spaCy + WordNet
• Adds academic transitions between sentences
• NEW: Adds hedging language to sound more natural and academic
• NEW: Intelligently combines short sentences with semantic analysis
• Configurable transformation probabilities for fine control

• Python 3.9+
• pip

1. Clone the repository:

git clone <repository-url>
cd Text-Humanizer-Python-Fork

2. Create a virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt
python -m spacy download en_core_web_sm
python -m nltk.downloader punkt punkt_tab wordnet averaged_perceptron_tagger

4. Run the API:

uvicorn main:app --reload

The API will be available at: http://localhost:8000

Humanize AI-generated text while preserving citations.

Request Body:

{
  "text": "Your AI-generated text here...",
  "synonym_probability": 0.2,
  "transition_probability": 0.2,
  "hedging_probability": 0.15,
  "sentence_combine_probability": 0.3
}

Response:

{
  "original_text": "...",
  "humanized_text": "...",
  "original_word_count": 150,
  "humanized_word_count": 165,
  "original_sentence_count": 8,
  "humanized_sentence_count": 8
}

Detect AI-generated content in text.

Request Body:

{
  "text": "Your text to analyze..."
}

Response:

{
  "text": "...",
  "classification_results": {
    "First sentence.": "AI-generated",
    "Second sentence.": "Human-written"
  },
  "percentages": {
    "AI-generated": 45.5,
    "AI-generated & AI-refined": 10.2,
    "Human-written": 30.1,
    "Human-written & AI-refined": 14.2
  },
  "summary": { ... }
}

Root endpoint with API information.

Health check endpoint.

Interactive API documentation (Swagger UI).

Humanize Text:

curl -X POST "http://localhost:8000/humanize" \
  -H "Content-Type: application/json" \
  -d '{
    "text": "AI is transforming industries. It is creating new opportunities.",
    "synonym_probability": 0.3,
    "transition_probability": 0.2,
    "hedging_probability": 0.15,
    "sentence_combine_probability": 0.3
  }'

Detect AI:

curl -X POST "http://localhost:8000/detect" \
  -H "Content-Type: application/json" \
  -d '{
    "text": "The research methodology employed a quantitative approach."
  }'

import requests

# Humanize text
response = requests.post(
    "http://localhost:8000/humanize",
    json={
        "text": "Your AI text here...",
        "synonym_probability": 0.2,
        "transition_probability": 0.2,
        "hedging_probability": 0.15,
        "sentence_combine_probability": 0.3
    }
)
result = response.json()
print(result['humanized_text'])

# Detect AI
response = requests.post(
    "http://localhost:8000/detect",
    json={"text": "Text to analyze..."}
)
result = response.json()
print(result['percentages'])

// Humanize text
fetch("http://localhost:8000/humanize", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    text: "Your AI text here...",
    synonym_probability: 0.2,
    transition_probability: 0.2,
  }),
})
  .then((res) => res.json())
  .then((data) => console.log(data));

// Detect AI
fetch("http://localhost:8000/detect", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    text: "Text to analyze...",
  }),
})
  .then((res) => res.json())
  .then((data) => console.log(data));

Run the test script to verify everything works:

python test_api.py

Or visit the interactive documentation at: http://localhost:8000/docs

Text-Humanizer-Python-Fork/
├── main.py                    # FastAPI application
├── utils/
│   ├── __init__.py
│   ├── ai_detection_utils.py  # AI detection logic
│   ├── model_loaders.py       # Model caching
│   └── text_humanizer.py      # Text humanization logic
├── requirements.txt           # Python dependencies
├── test_api.py                # Test script
├── .gitignore                 # Git ignore rules
└── README.md                  # This file

• FastAPI - Modern web framework for building APIs
• Transformers - HuggingFace library for AI models
• spaCy - Industrial-strength NLP
• NLTK - Natural Language Toolkit
• PyTorch - Deep learning framework
• Uvicorn - ASGI server

| Parameter | Type | Default | Range | Description