The Complex Langevin method is used to circumvent the sign problem in Monte Carlo simulations. I will review recent results for QCD, as well as recent theoretical developments. First the boundary terms will be reviewed, which signal when complex Langevin results are to be trusted. I will also show that in certain cases (e.g. when using appropriate kernels), the results can be understood as integrals over different integration cycles available in the analytically continued theory. For QCD, I show results for the equation of state at high temperatures, as well as results for dynamical stabilization, which might allow the exploration of the low temperature part of the QCD phase diagram. Dynamical stabilization can be understood as soft cutoff in imaginary directions of the complexified manifold. The bias caused by the cutoff can be removed to a good accuracy using extrapolation to infinite cutoff.