The trick is you must use the web api or you do it in Tier2.
I tried it for you and got an api information response
then i tried to get a file from the Web api to save and play it
but i had no api credits anymore...
So I believe it is possible but you must pay for the credits if you want to use Tier1.
If you need more information just ask ; )
[Edit]
Text to speech will work.
but speech recognition will be harder.
You cant read from the mic directly so you have to wait till the recorded file is saved.
You can make a stop record button and then transfer the file in the right format to the site and such a hassle.
[/Edit]