Plagiarism screening for Ayurveda
Built for PG, PhD, and journal workflows where generic detectors drown in shared dravya names and miss paraphrased ślokas. This release runs domain-tuned scoring via FastAPI with embeddings, cosine similarity, fuzzy matching, and Ayurvedic keyword boosting.
Paste text below. Screening is assistive not a misconduct finding.
Typical response in ~1.5–6s
Longer excerpts yield stronger segment-level analysis.
URL ingestion is on the roadmap — paste article text for now.