[alex_kac]

I am looking for a server type application that will take speech pre-recorded in WAV files and convert them into text. I would really like to have it run on OS X, but would be OK with pretty much any system including Windows.

I have looked at Nuance's Speech Server but they want $25k for it + $2k just to test it. I am also looking at the open source Sphinx4. I am hoping someone here might know of something that would work.

[Mithras]

In the commercial realm, I don't think you'll do much better than using Nuance's DNS with some scripts to feed it. The server product will offer better throughput and scriptability, but depending on your needs, the regular home version should work OK.

You can also tinker with CMU's Sphinx, but recognize that the publicly available corpora are not as large or as mature as the commercial ones.