*Phaser* is a software program designed for phasing macromolecular crystal structures using both molecular replacement and experimental phasing methods. It employs novel phasing algorithms based on maximum likelihood and multivariate statistics, which have been developed to improve the discrimination of correct solutions from noise compared to traditional methods. *Phaser* is highly automated, allowing for direct scripting from Python and traditional *CCP4* keyword-style input. The software aims to facilitate the development and release of improved phasing methods to the crystallographic community. The algorithms in *Phaser* include maximum likelihood phasing for rotation and translation functions in molecular replacement and the SAD function in experimental phasing. These algorithms incorporate experimental error and account for correlations between structure factors, enhancing the discrimination of correct solutions. The software also includes anisotropy correction to handle systematic variations in reflection intensities and fast rotation and translation functions to improve computational efficiency. Additionally, *Phaser* incorporates multivariate statistics to handle correlations between structure factors, particularly in single-wavelength anomalous dispersion (SAD) phasing. It also features an ensemble procedure to generate a single calculated structure factor set from a set of structurally aligned models. Furthermore, *Phaser* supports normal-mode analysis to explore conformational changes in models, which can help generate models closer to the target structure. Overall, *Phaser* is designed to be a versatile and powerful tool for crystallographic phasing, offering both improved automation and advanced algorithms to enhance the success rate of structure solution.*Phaser* is a software program designed for phasing macromolecular crystal structures using both molecular replacement and experimental phasing methods. It employs novel phasing algorithms based on maximum likelihood and multivariate statistics, which have been developed to improve the discrimination of correct solutions from noise compared to traditional methods. *Phaser* is highly automated, allowing for direct scripting from Python and traditional *CCP4* keyword-style input. The software aims to facilitate the development and release of improved phasing methods to the crystallographic community. The algorithms in *Phaser* include maximum likelihood phasing for rotation and translation functions in molecular replacement and the SAD function in experimental phasing. These algorithms incorporate experimental error and account for correlations between structure factors, enhancing the discrimination of correct solutions. The software also includes anisotropy correction to handle systematic variations in reflection intensities and fast rotation and translation functions to improve computational efficiency. Additionally, *Phaser* incorporates multivariate statistics to handle correlations between structure factors, particularly in single-wavelength anomalous dispersion (SAD) phasing. It also features an ensemble procedure to generate a single calculated structure factor set from a set of structurally aligned models. Furthermore, *Phaser* supports normal-mode analysis to explore conformational changes in models, which can help generate models closer to the target structure. Overall, *Phaser* is designed to be a versatile and powerful tool for crystallographic phasing, offering both improved automation and advanced algorithms to enhance the success rate of structure solution.