The only output format (right now) for an REST API call is to save it to an CSV file. This CSV file can be used to load an temporary cube.
A direct load (without saving the file to disk) into a cube / temporary cube would be a great enhancement.
I would also like to highlight the request for this feature.
This especially would help with parallel executions as a temporary cube automatically is only valid for each execution.
Currently I have the problem with parallelity as I face a scenario where the procedure with the API call step in it is also called with an API. This leads to the fact that I can not user "@User" in the filename to distinct the files because this will always result in the username of the API user.
A substitution formula like "@[entityname]" would work for the output filename of the API call step but this is not supported by the required Data Reader in the file path to read the file.