FYI. word_info v0.2.0 has been published, changelogs:

• some improvements on performance.
  In the first released version, dictionary data is compiled directly into BEAM codes. This may cause freezing (~3s) on starting the application, which may be a problem for an application who needs to boot as fast as possible.

  In v0.2.0, the data is converted into an ETS table dump at compile-time. This allows the code to be tidy and provides a fast boot.

• fix a typo on returning value

PS. I want to thank all the people and their valuable discussions involved in this thread, which leads me to a better approach:

I do not actually need this right now, but I am sure that if someone has a need for such a thing, it might also be needed for other languages other than english.

That’s what I was going to say as well. I would like to be able to use such a library, but I would consider it only if it works for the different languages supported on my website.

It would be super nice to get the IPA pronunciation of words in many different languages as well. Arpabet seems to be restricted to English so that function would only work for that language I guess.

I think both pronunciation and syllables can be derived programmatically for quite a few languages (easy mode for Spanish/Italian/Korean), if no extensive resource exists on the net.