Integrating pywhispercpp as the first extension to lollms-webui

[ParisNeo]

Hi Abdeladim. I finally start to write extensions to lollms and I was thinking that first extension should be audio in and audio out. But I need to comply with my rule number 1: Every thing should be done locally. No data is sent anywhere out of your PC.

To do this, I think whisper is really cool. Even cooler is whispercpp, but since I use python, i need pywhispercpp :)

Do you have an example of your code that uses direct input stream from microphone? That would simplify the integration greatly.

[absadiki]

Hi Saifeddine,

That's great to hear. We'll finally see lollms with audio :)

Yes I have done an example that uses direct input stream from the mic, it is the assistant example.
I created it in this way because I had the idea to deploy it in a raspberry PI as an assistant.

The class Assistant is already ready to be imported and you only need to pass a callback to it. Otherwise you can tweak the source code to serve your needs. It's fairly simple.

Let me know if you need anything else.

[ParisNeo]

Wonderful!