We develop a new R package, EAinference, for statistical inference on high-dimensional data. Methods in this package are implemented based on the idea of estimator augmentation (Zhou 2014 and Zhou and Min 2017), which can be used to calculate the joint density of a regularized estimator and its subgradient in the KKT conditions. Our package implements estimator augmentation for many popular sparse estimators, including the lasso, the group lasso, and the scaled lasso among others. It provides methods to simulate from the sampling distribution of a sparse estimator via parametric bootstrap, importance sampling and Markov chain Monte Carlo (MCMC) for high-dimensional inference. In particular, we propose a new method for post-selection inference with the lasso estimator, using an MCMC sampler to draw from the conditional distribution of the lasso given a selected model. Applied to extensive simulated data, our method produces shorter confidence intervals with accurate coverage rate and shows a much higher power for detecting active variables, compared to a state-of-the-art method.