Yes, you can use cxcalc to predict the microspecies distribution of your molecules. All functions that are available in Marvin can be accessed via cxcalc as well.
However, you can't use in-house pKa data to predict microspecies distribution directly (that is you can't predict microspecies data solely based on your pKa values). The only way to do this is to train our predictor with your in-house pKa data. If the training is successful, your data will be taken into account.
I hope this helps,