Is it possible to automate the voice streaming using AI tools?

[Azhar Ali]

Hi,

We use flashphoner server to broadcast content to our listeners using webrtc. We are interested to know if it's possible to use AI text to voice tools and auto stream the generate voice. There are many tools out there with API which will return from mp3 or some format of voice stream.

Users are already connected to the stream which was streaming using a mic, after a user has finished talking, we want the AI to take over and read content which is generated by tools. We don't want to disconnect the users, ideally same stream

https://www.liveperson.com/
https://speechify.com/
https://lovo.ai/text-to-speech
https://www.unite.ai/goto/murf
https://synthesys.io/

Anything like this possible?

 

[Max]

Good day.
This seems to be an integration task:
1. A TTS system should generate an audio file
2. The file should be converted to mp4 container and placed to /usr/local/FlashphonerWebCallServer/media (or a folder defined by media_dir parameter)
3. Stream injection feature should be used (enable /stream/inject2 API by use_new_injector=true parameter):

POST /stream/inject2/startup HTTP/1.1
Host: localhost:8081
Content-Type: application/json

{
    "localStreamName":"stream",
    "remoteStreamName":"vod-live://tts-file.mp4",
    "video":false,
    "audio":true
}

Please read details here: Injection implementation in build 5.2.1618 and newer