A downloadable tool for Windows

Description : 

VCompanionPy is a small application that uses a combination of python packages ( coqui TTS, twitchio, OpenAI API, fast whisper, eel, Pyaudio..) to give you a streaming companion! 

However the app is using the API of OpenAI which means you will  have to use your own OpenAI key. 

But what is a streaming companion ? Well, it's an AI that can respond to your chat (vocally using a TTS and textually using a .txt file that you can add as subtitles on OBS) or talk with you! (if the vocal conversation option is activated : the AI will listen to your mic and answer you!)

You can choose the audio output of the TTS which means you can use a third party software such as Vtube Studio to give the AI a "virtual" body!

The TTS also has a wide variety of voices (108 to be exact) that you can try out and use for free!

As for the TTS : You have access to ALL the free TTS given by Coqui TTS! Check out their github page for more information on their TTS list (or try them in the app!).

The only thing that needs money in this app is the OpenAI API! (however if you want to buy me a coffee on patreon i'll be very grateful ^_^ : https://akt0o.carrd.co/)

Important :

This app is in its early stages! Any feedback is greatly appreciated! Also I haven't had a chance to test the OpenAI API so any feedback on how well the AI works is appreciated!

If you find a bug or have a problem : don't hesitate to comment in this page!

PS : License in the files ^^

More information

Download

This tool is currently unavailable

Development log

Comments

Log in with itch.io to leave a comment.

Akt0o261 days ago

if this application draws enough interest, I have these features in mind that I could add : 

-Access to all the free TTS of coqui TTS (which mean : TTS for other language than english) + the possibility to connect to your coqui TTS account to use the paid TTS they offer.

-A version that uses a local AI (a 7b model or something like that, for reference : a 7b AI can run on around 8gb of ram or vram).  This would mean : no need to pay at all, but slower text generation from the AI and a much more resource consuming app. (YonmaAI is a local AI if you want an example, she takes between 20seconds and 50seconds to generate text)

-Any suggestions?

Reply
Akt0o258 days ago

access to all free tts of coqui TTS : done

Reply