Azure Kinect, Realsense, ZED all have support for recording sessions to files.
It would be nice to be able to use a recording as a source to depthkit, instead of live sensor.
This way you can record on different machine that the one you do the processing on. This help ease logistics of field recording.
You can also pre-process the file sources using the relevant SDKs and provide that file to Depthkit. This would be useful if you are using a multiple sensor setup.