The MR-Base platform supports systematic causal inference across the human phenome by integrating a curated database of complete genome-wide association study (GWAS) results with an application programming interface, web app, and R packages that automate 2-sample Mendelian randomization (2SMR). This platform enables the efficient evaluation of millions of potential causal relationships in phenome-wide association studies (PheWAS) by providing sensitivity analyses for assessing horizontal pleiotropy and other violations of assumptions. The database currently contains 11 billion single nucleotide polymorphism (SNP)-trait associations from 1673 GWAS studies and is regularly updated. MR-Base automates the process of 2SMR, making research faster and more reliable, and allows researchers to investigate potential causal relationships between traits using a wide range of statistical methods. The platform supports hypothesis-driven analyses and hypothesis-free MR-PheWAS to identify potential sources of horizontal pleiotropy and gain insights into the impacts of interventions. MR-Base also facilitates the integration of data and analytical tools, enabling novel insights that would have been technically and practically challenging to achieve. The platform includes several sensitivity analyses to account for potential patterns of horizontal pleiotropy and provides a range of statistical methods for causal inference, including inverse variance weighted (IVW) meta analysis, maximum likelihood, MR Egger analysis, median-based estimator, and mode-based methods. The results of MR-Base analyses are presented alongside sensitivity analyses to ensure reliability and reproducibility. The platform is designed to support systematic causal inference across the human phenome by integrating GWAS summary data with statistical methods for causal inference.The MR-Base platform supports systematic causal inference across the human phenome by integrating a curated database of complete genome-wide association study (GWAS) results with an application programming interface, web app, and R packages that automate 2-sample Mendelian randomization (2SMR). This platform enables the efficient evaluation of millions of potential causal relationships in phenome-wide association studies (PheWAS) by providing sensitivity analyses for assessing horizontal pleiotropy and other violations of assumptions. The database currently contains 11 billion single nucleotide polymorphism (SNP)-trait associations from 1673 GWAS studies and is regularly updated. MR-Base automates the process of 2SMR, making research faster and more reliable, and allows researchers to investigate potential causal relationships between traits using a wide range of statistical methods. The platform supports hypothesis-driven analyses and hypothesis-free MR-PheWAS to identify potential sources of horizontal pleiotropy and gain insights into the impacts of interventions. MR-Base also facilitates the integration of data and analytical tools, enabling novel insights that would have been technically and practically challenging to achieve. The platform includes several sensitivity analyses to account for potential patterns of horizontal pleiotropy and provides a range of statistical methods for causal inference, including inverse variance weighted (IVW) meta analysis, maximum likelihood, MR Egger analysis, median-based estimator, and mode-based methods. The results of MR-Base analyses are presented alongside sensitivity analyses to ensure reliability and reproducibility. The platform is designed to support systematic causal inference across the human phenome by integrating GWAS summary data with statistical methods for causal inference.