If you are broadcasting wikipedia does this mean somewhere in the UI however it’s deployed I’ll have a search box and can search for any article in wikipedia when I need it?
That’s still not decided. Most likely, we’ll harvest only content (Wikipedia is written in PHP, but we won’t broadcast any PHP code). I think an optimal solution would be for the client software (one on the user’s computer) to build a local index each time it fetches new content and make the local archive searchable.
Kiwix (http://www.kiwix.org) provides a range of solutions to do this. TED talks, Wikipedia or Gutenberg are already packaged in heavy compressed and directly readable files (the ZIM files http://www.openzim.org). We would love to help you to choose the best approach to spread efficiently the content.