The kiwix project helps people download and search data offline:
One facility at the project enables users to download a dump of Wikipedia, along with an index, which means people can keep and search the whole of Wikipedia without using the internet.
Might YaCy be able to make use of the Kiwix indexes of Wikipedias (they are available in several languages.)
Would use of these indexes be able to reduce redundant crawls of Wikipedia amongst different peers?
Apart from Wikipedia, Kiwix supports several other data projects that might be of use in searches.