Just go to
skrutable.pythonanywhere.com, enter (e.g. copy and paste) some Sanskrit text into the upper box, check the transliteration settings, and hit the button "Split Sandhi & Cpds" (i.e., “compounds”). After a brief wait, the output appears in the lower box, with new spaces in place of both dissolved inter-word and intra-compound sandhi. Punctuation is (mostly) preserved, and you (mostly) don't need to worry about length limits. There's also a whole-file option which is relatively fast.
To clarify, whereas Skrutable's transliteration and meter tools, available via the same interface, are programmed by myself, I haven't changed or improved the 2018 Splitter tool at all, but rather just facilitated access to it with the authors' blessings. As they themselves point out, eventually this Splitter tool should be improved and/or superseded, e.g. in order to deal more robustly with various orthographies and genre-specific idioms, and/or to distinguish between inter-word and intra-compound boundaries, but for now, even with less-than-perfectly ideal output, it's still quite nice to have on hand, as I think you'll now be able to agree.
Kudos and thanks again to Oliver and Sebastian for their great work. I'll happily consider feedback on the interface in particular.
Kind regards,
Tyler