I was wondering what name should be given to the Speech-to-text API key? In the previous version, the format for service private keys on Pavlovia looks different than the current format. This could be the reason that I can’t use the API key. Should I just give a random name to it?
I checked the source code. I believe the key should be read by the Pavlovia query through the ServerManager code from PsychoJS below. However, I use console.log() to debug and found it actually didn’t correctly read anything in the “Service Private Keys” block.
I know there’s a official demo : transcription-demo [PsychoPy] which used Google-Cloud Speech-to-text API correctly.
Plus, I compared my JS file and the JS file extract from the back-end, there’s no difference in calling the engine.GOOGLE. However, I will receive a error: missing key for the engine.
solved
What was your solution?
I didn’t really see the transcription, so when I solve the routine skip problem I posted yesterday, I can know wether the google cloud really works.
My solution is to put the API key in the code (not safe), and write a similar code like in psychoJS to combine transcribe(), and _GoogleTranscribe() but avoid using them, because these functions read the key from the system data. Still debugging but it works outside of Pavlovia.
Okay… It seems I still can’t make it work with PsychoPy online experiment