Paste a Bangla news article or a link. The system compares it against trusted reporting and returns an authenticity score built from four signals — knowledge consistency, semantic similarity, sentiment, and source credibility.
The system will crawl and read the article, then run detection.
Try a sample:
Live scoring runs through the PHP backend → Python NLP service (NER, Doc2Vec, Bengali VADER, source credibility). If the Python service is offline, PHP uses a transparent fallback so results still appear.