The R package 'dagitty' provides a powerful tool for robust causal inference using directed acyclic graphs (DAGs). It allows users to analyze and draw DAGs within the R statistical environment, offering functionalities similar to the web-based DAGitty application. The package enables evaluation of whether a DAG is consistent with the dataset it represents, enumeration of statistically equivalent but causally different DAGs, and identification of valid adjustment sets for different DAGs. These features help epidemiologists detect causal misspecifications and make robust inferences. DAGs are used to represent causal relationships and identify covariate adjustment sets to minimize confounding bias. However, applying graphical criteria like the back-door criterion becomes cumbersome with larger DAGs. The web application DAGitty was developed to address this, but integrating its functionality into statistical software like R enhances usability. The R package 'dagitty' bridges this gap by providing access to DAGitty's features within R, along with additional functions. The package uses a library of JavaScript routines integrated via the 'V8' package, allowing it to maintain synchronization with the web application. It supports both linear and non-linear regression for testing conditional independence, and offers functions to evaluate DAG-dataset consistency. For example, it can test whether implied independencies hold in the dataset, helping to identify potential errors in DAG specification. The R package 'dagitty' also allows users to identify adjustment sets that are valid across statistically equivalent DAGs. This is particularly useful as different DAGs can have the same testable implications. The package includes functions to generate equivalent DAGs and determine adjustment sets for these, ensuring robustness against potential misspecifications. The package is available on CRAN and GitHub, and the web application is free software under the GPL license. It supports a wide range of analyses, including evaluating the consistency of DAGs with datasets, identifying adjustment sets, and testing for conditional independence. The package helps address concerns about the validity of DAG-based analyses by providing tools to evaluate assumptions against dataset implications, thus enhancing confidence in causal inferences.The R package 'dagitty' provides a powerful tool for robust causal inference using directed acyclic graphs (DAGs). It allows users to analyze and draw DAGs within the R statistical environment, offering functionalities similar to the web-based DAGitty application. The package enables evaluation of whether a DAG is consistent with the dataset it represents, enumeration of statistically equivalent but causally different DAGs, and identification of valid adjustment sets for different DAGs. These features help epidemiologists detect causal misspecifications and make robust inferences. DAGs are used to represent causal relationships and identify covariate adjustment sets to minimize confounding bias. However, applying graphical criteria like the back-door criterion becomes cumbersome with larger DAGs. The web application DAGitty was developed to address this, but integrating its functionality into statistical software like R enhances usability. The R package 'dagitty' bridges this gap by providing access to DAGitty's features within R, along with additional functions. The package uses a library of JavaScript routines integrated via the 'V8' package, allowing it to maintain synchronization with the web application. It supports both linear and non-linear regression for testing conditional independence, and offers functions to evaluate DAG-dataset consistency. For example, it can test whether implied independencies hold in the dataset, helping to identify potential errors in DAG specification. The R package 'dagitty' also allows users to identify adjustment sets that are valid across statistically equivalent DAGs. This is particularly useful as different DAGs can have the same testable implications. The package includes functions to generate equivalent DAGs and determine adjustment sets for these, ensuring robustness against potential misspecifications. The package is available on CRAN and GitHub, and the web application is free software under the GPL license. It supports a wide range of analyses, including evaluating the consistency of DAGs with datasets, identifying adjustment sets, and testing for conditional independence. The package helps address concerns about the validity of DAG-based analyses by providing tools to evaluate assumptions against dataset implications, thus enhancing confidence in causal inferences.