ARKLab provides a suite of Matlab files that implement adaptive-step explicit, implicit and additive Runge-Kutta methods. This software is designed for instructional use and has been implemented to be easily understood and modified. As a result, certain efficiency-related optimizations have been omitted. We currently provide three classes of methods: (a) explicit Runge-Kutta (ERK), (b) diagonally-implicit Runge-Kutta (DIRK), and (c) additive Runge-Kutta (ARK). In addition to these three classes of methods, we provide a large number of Butcher tables (over 70), holding the coefficients for existing methods in each of these three categories. For these methods, time step adaptivity is enabled for methods with embeddings. All three classes of methods may be run in so-called "fixed-step mode", wherein the solver will take steps of a user-supplied magnitude. This mode is currently required for all methods without embeddings. In addition to these sets of solvers, we provide a set of simple example problems that may be used to test different methods, and that may be used as a template for creating new problems. These test problems are described in the accompanying file, 'Problems.txt'.