Dechecker's AI Checker and Detector Tool

• English | 中文

Dechecker is a powerful AI Checker tool designed to detect whether a piece of text was written by humans or generated by AI models like ChatGPT, GPT-5, Claude, Gemini, or LLaMA. It provides sentence-level AI detection, highlights AI-written parts, and helps users optimize text for authenticity and originality.

This repository contains the backend Python project of Dechecker’s AI Checker, offering a RESTful API for text detection and analysis. You can integrate it into your web app, Chrome extension, or internal moderation tools.

👉 Live product reference: https://dechecker.ai

• AI Detection – Identify whether text is AI-generated or human-written with confidence scores.
• Sentence-Level Analysis – Highlight specific sentences likely generated by AI.
• Content Authenticity Score – Get a numerical score representing the probability of AI authorship.
• Optimization Suggestions – (Optional) Rewrite or refine AI-written content to sound more human and natural.
• Multi-use Scenarios – Perfect for academic papers, SEO content, blogs, social media posts, essays, and copywriting.
• Fast & Lightweight API – Built with FastAPI/Flask for easy deployment and integration.

• Language: Python 3.10+
• Framework: FastAPI (or Flask)
• Model Backend: Hugging Face Transformers, PyTorch
• AI Detection Model: Fine-tuned RoBERTa (example)
• Environment: Docker-ready, REST API
• License: MIT

git clone https://github.com/yourusername/dechecker-ai-checker-backend.git
cd dechecker-ai-checker-backend

pip install -r requirements.txt

Create a .env file or export directly:

API_KEY=your_api_key_here
MAX_TEXT_LENGTH=20000
MODEL_PATH=/path/to/your/model

If using FastAPI:

uvicorn app.main:app --reload

If using Flask:

python app.py

Detect whether a text is AI-generated or human-written.

Request

POST /api/detect
Content-Type: application/json

{
  "text": "This is a sample text written for AI detection."
}

Response

{
  "score": 0.87,
  "is_ai_generated": true,
  "analysis": [
    {
      "offset": 35,
      "length": 42,
      "snippet": "This section is likely AI-generated.",
      "confidence": 0.93
    }
  ]
}

Generate rewrite suggestions to improve text originality.

Request

POST /api/suggest
Content-Type: application/json

{
  "text": "This content may sound AI-generated."
}

Response

{
  "suggestions": [
    "Consider adding a personal opinion or anecdote.",
    "Use more varied sentence structures to enhance flow."
  ]
}

import requests

url = "http://localhost:8000/api/detect"
payload = {"text": "AI-generated content detection is essential."}
resp = requests.post(url, json=payload)
data = resp.json()

print("AI Probability:", data["score"])
print("AI Detected:", data["is_ai_generated"])
for seg in data["analysis"]:
    print(f"Snippet: {seg['snippet']} (Confidence: {seg['confidence']:.2f})")

• Docker Support: Easily deploy Dechecker’s AI Checker backend using Docker.
• Scalability: Supports async requests and horizontal scaling via Kubernetes or Docker Compose.
• Security: Add authentication (API keys or OAuth) to prevent abuse.
• Queue Processing: Use Celery or RabbitMQ for batch text detection jobs.
• Monitoring: Integrate Prometheus or ELK for observability.

We welcome contributions! Please open an issue or submit a pull request with improvements, bug fixes, or feature suggestions.

Before submitting:

• Follow PEP8 coding standards.
• Include clear commit messages.
• Run tests and lint checks.

• 🌐 Official Site: Dechecker
• 📘 API Docs: Coming soon
• 🧑‍💻 Developer Contact: support@dechecker.ai