Both biahub and autotracker projects require the ability to virtually stain 3D stacks via a python method that implements the steps of sliding window inference and blending. This method should be practical to parallelize and use during acquisitions.