Scheduled extractions resemble what was done with TRTH v1. In TRTH v2 they can be done using the GUI or the API. They require an instrument list, report template and schedule, all saved on the server. The schedule can be recurring or occur a single time, in the future or immediately.

On Demand extractions are done on the fly, and can only be requested using the API. The API call specifies the data type, and the data fields are passed to the call as parameters, as well as the instrument list. Nothing is saved on the server. Consider these calls as high level simplified calls.

You can use any one you like, to best suit your use cases. Typically you'd use scheduled extractions for regular extractions, and on demand extractions for ad hoc requests.