The MCMCglmm R package implements Markov chain Monte Carlo (MCMC) methods for fitting multi-response generalized linear mixed models (GLMMs). These models are flexible and can handle non-Gaussian response variables, which cannot be modeled using traditional maximum likelihood methods. MCMCglmm allows for multiple response variables, including Gaussian, Poisson, multinomial, exponential, zero-inflated, and censored distributions. It supports a wide range of variance structures for random effects and residuals, including interactions with categorical or continuous variables and complex structures arising from shared ancestry (e.g., pedigree or phylogeny). The package uses C/C++ for efficient computation, leveraging the CSparse library for sparse linear systems. MCMCglmm is designed to be user-friendly, requiring minimal expertise while reducing computational time. It is compared to WinBUGS, where it is found to be approximately 40 times faster per iteration and has a higher effective sample size per iteration. The package is particularly useful for quantitative genetic analyses, where it can model complex relationships between traits and individuals. It supports a variety of distributions and variance structures, making it suitable for a wide range of applications. The package is available from CRAN and is used for analyzing data with multiple response variables, including examples from quantitative genetics and meta-analysis. The MCMCglmm package provides efficient and flexible tools for fitting GLMMs, with a focus on computational efficiency and ease of use.The MCMCglmm R package implements Markov chain Monte Carlo (MCMC) methods for fitting multi-response generalized linear mixed models (GLMMs). These models are flexible and can handle non-Gaussian response variables, which cannot be modeled using traditional maximum likelihood methods. MCMCglmm allows for multiple response variables, including Gaussian, Poisson, multinomial, exponential, zero-inflated, and censored distributions. It supports a wide range of variance structures for random effects and residuals, including interactions with categorical or continuous variables and complex structures arising from shared ancestry (e.g., pedigree or phylogeny). The package uses C/C++ for efficient computation, leveraging the CSparse library for sparse linear systems. MCMCglmm is designed to be user-friendly, requiring minimal expertise while reducing computational time. It is compared to WinBUGS, where it is found to be approximately 40 times faster per iteration and has a higher effective sample size per iteration. The package is particularly useful for quantitative genetic analyses, where it can model complex relationships between traits and individuals. It supports a variety of distributions and variance structures, making it suitable for a wide range of applications. The package is available from CRAN and is used for analyzing data with multiple response variables, including examples from quantitative genetics and meta-analysis. The MCMCglmm package provides efficient and flexible tools for fitting GLMMs, with a focus on computational efficiency and ease of use.