Many people use Amazon's mechanical Turk for this. It allows you to make automated requests like this that are processed by real people. Perhaps it is not the easiest service to dive into, so I wonder if their terms of service would allow someone to make an easy to use front end that focuses on just this task.
Comments
Speech recognition: send an audio file, receive text.
AT&T offers this: http://developer.att.com/developer/forward.jsp?passedItemId=...
Many people use Amazon's mechanical Turk for this. It allows you to make automated requests like this that are processed by real people. Perhaps it is not the easiest service to dive into, so I wonder if their terms of service would allow someone to make an easy to use front end that focuses on just this task.
Check out http://speakertext.com/