Make sure that enough functionality is kept to enable recitation without javascript. This means it should still be possible to preprocess text in onParserAfterTidy(), but some information, like utterance offsets, shouldn't be needed.
The desired output likely has significant overlaps with that we will want in the Special page for downloading all the audio of an article.
This task includes a proof of concept solution for surfacing the output to a no-js user. (Link for playing utterance next to each sentence).
Properly enabling it for no-js scenarios is a follow-up task (to be created).