I wrote a search tool to improve on Grep, which I currently use to look for parallels in all the available E-texts (Gretil, Muktabodha, Sarit and Titus combined). The tool takes care of various encoding issues, notational conventions and oddities commonly found in Sanskrit E-texts. It also knows about sandhi (e.g. looking for "smṛtes" will match "smṛteḥ" and "smṛter api", see the screenshot below).
The tool is written in such a way that missing a textual parallel is extremely less likely than with Grep and such. However, it tends to
over-generate matches, which I believe is an acceptable trade-off when searching for textual parallels. Contrarily to Prof. Kulkarni's Gaveṣika, it doesn't use a parser, which makes it resilient to "difficult" or corrupt texts, typos, etc.
I haven't made the tool public so far, because this would require a lot of additional work on my part. I however intend to do it as soon as I find enough time.
Best regards,
Michaël Meyer