The Model Card Toolkit facilitates the generation of “model cards,” a concept first introduced in the paper “Model Cards for Model Reporting.” This paper established the first-pass set of transparency- and fairness-oriented documentation procedures. Model cards include sections on the intended use case, metrics, evaluation and training data, and intersectional quantitative analysis. For machine learning pipelines that leverage the Tensorflow Extended (TFX) platform or the ML Metadata library, this project automates the generation of model cards, whose schemas are stored in JSON as an intermediate format.