edit: I get it, OP just keeps submitting his service with different descriptions until one gets some upvotes. Only took 25 tries to get 30 points. Shameful.
From my experience with NLP/AST the tricky part is models for some less common languages.
Clicking on the Sign Up button on iOS Safari does nothing.
Clicking on the Get Started button takes me to an Upload Video form - not what I expected from a mp3-to-text service.
Here's an example using GNU parallel: http://voice2json.org/recipes.html#parallel-wav-recognition
> Sets of voice commands that are described well by a grammar
> Commands with uncommon words or pronunciations
> Commands or intents that can vary at runtime
Doesn't sound like what you'd want for a generic transcription service.