The only possibility I have found is setting CMUSphinx as a web server
Google speech API is neither unlimited(1) nor safe to implement.
ATT Speech to Text is free but the app keys expire after 90 days and must be renewed(2), and a lower maximum number of transactions allowed per day(3)
I think that setting CMUSphinx as a web server would take a long time, so it is wise to consider other options. Do you have some recommendation?
Note: we are using Android, but the Android API only allows to recognize audio from the microphone, not from an audio file.
1) When you obtain a free key, you only get a quota of 50 calls to the service per day:
2) I'm not sure if that means if we must compile our software again.
3) I didn't find what is that limit exactly