Stylometric
Deep Learning
Arabic NLP
Arabic AI Detector
Paste any Arabic article below. Our model uses stylometric features and a fine-tuned Transformer to classify whether the text was written by a human or generated by AI.
How It Works
01
Text Cleaning
The raw text is cleaned with a noise-removal pipeline: HTML tags, diacritics, and non-Arabic/English tokens are stripped.
02
Stylometric Extraction
32 stylometric features are extracted to build a precise numeric fingerprint of the writing style.
03
Deep Learning Classification
A fine-tuned AraELECTRA Transformer + stylometric features are fed into a hybrid classifier trained on thousands of Arabic articles.