I am looking for someone who can set up an open-source speech-to-text server for me.
Which can convert all mp3 recordings of my organization from speech to text.
It is not important to convert to complete sentences, but you can convert word to word.
The reason for this solution is I want to detect bad words in the converted file. I will create another API that will check bad words in this (converted) text file.
My organization speaks English, Urdu, Arabic (all mixed within the same recording).
That is why I don't want to convert to sentences, instead convert word to word.
after I award the project to you and before we start our work, I will give you a few MP3 samples of recordings. You will need to process them and give me a text file. So I can check the accuracy of speech-to-text of the server.
Once we finalize a server, You can proceed to set up the server and APIs.
I have around 100 recordings of 30 minutes each, every hour in my office hours.
So the system should work faster while consuming the normal CPU and RAM of my dedicated server.
The accuracy of this should be higher.