ml‑knn is a lightweight JavaScript library that implements a general‑purpose k‑nearest‑neighbors classifier using a kd‑tree for fast neighbor searches. It lets developers train a model with a numeric dataset and label array, then predict labels for new samples, and supports custom distance functions and model serialization. The package is aimed at web or Node.js developers who need an easy‑to‑integrate, pure‑JS machine‑learning classifier.