I have a large number of MHT files on my system, created by saving a webpage in my web browser (the majority saved by Firefox using the UnMHT add-on, but some saved by Internet Explorer or Opera). Though DocFetcher seems to find words in these files (they are readable text after all, except for images etc which are encoded), when you click on a search match in DocFetcher, nothing is shown in the preview. That means you have to double click on each match to open it in your web browser, then use Ctrl+F in the browser to find the search term. This is not really practical when DocFetcher finds hundreds of MHT files in a set of search results.
So it would be great if DocFetcher could fully support MHT files. The RFC for MHT (technically MHTML) files is linked in the Wikipedia article: