Model predictive control applications have broadened from industrial plants with slow system dynamics to the implementation of control algorithms on low-cost hardware. Ip2go has been developed with the aim of making model predictive control readily accessible on embedded systems. It generates efficient, problem-tailored solvers for optimal control problems. The code generator implements Mehrotra’s predictor-corrector interior-point method with Riccati recursions. It can be applied to quadratic programming problems with linear discrete-time system dynamics and polytopic inequality constraints. Both hard and soft constraints can be implemented. In this paper the features of ip2go, the underlying algorithm, the benefits of code generation and the generation process are set out. The comparison of ip2go-generated solvers to state-of-the-art solvers shows that ip2go can compete in terms of efficiency and code-size. The open and modular design allows all users to amend the generator. The Matlab and based and LGPL licensed C-code generator ip2go is available under https://github.com/fabiankrank/ip2go.