Develop API calls and pushes to other systems in-browser
Actions are used in workflows and can do anything from sending data, to triggering events, to fetching records. Workflows are capable of accommodating a chain of actions. This means that you could have an action that fetched data crucial to your workflow, that action can be reference later in the workflow. Actions are especially flexible and can even be used to perform functions internal to the workflow, such as data manipulation. The three components of an action are the input object, the output object, and the Python code to perform the action and yield the response.

Input object

All actions require an input object. The input object is essentially a form that the Middle user will need to populate with static and variable data on the workflow side. Setting up the input object is similar to creating record type attributes: you create a field, select its data type, input the destination's internal field name, name it, and provide a description. It's always good to note any intricacies regarding a data field in the description. Input objects can accomodate embedded arrays of objects.

Output object

The output object is setup similar to the input object, however this is the set of data that will be accessible as a global variable in the workflow after the action is performed. The output object is typically populated by the response from the destination system. Output objects can include anything from success or error messages to the data returned by the action, such as a fetched user record or a record that was created by the action.
Configuring an output object in an action