›KhipuKoding‹ is a sound performance thought around the »Khipu«, a fiber-based notation device of pre-Hispanic origin, that combines Machine Learning, Computer Vision, the human act of knotting, and a hybrid coding method that reads khipu-style knots in order to generate sound. The KhipuKoding system consists of hanging textile ropes that are tied following a notation system inspired in the actual values given by the Incan khipu method. A computer running a Machine Learning + Computer Vision model (Yolo5) trained to recognize knots then associates their position to numbers (in the Incan manner) and that, in turn, runs or modifies a line of code in the FoxDot (Supercollider) environment in order to generate sound.