There are various sources of Sanskrit linguistic data available online, but they vary widely in coverage and quality. So I've spent some time collating the best of this data into a simple format that's easy to use.
There are still some problems to work out, but I think that this data is useful enough right now that it's worth sharing:
https://github.com/sanskrit/data
Please report any data issues either here or on the GitHub page.