Tensor trains represent high-rank tensors as chains of small cores, turning exponentially large arrays into a compact format that supports decomposition, compression, and arithmetic. The functions fall into a few natural groups: construction, conversion to and from dense arrays, rank control and arithmetic.