search

nomic-ai / gpt4all

GPT4All: Chat with Local LLMs on Any Device
https://gpt4all.io
MIT License
67.71k stars 7.44k forks source link

Bark (AI-generated voice output) #806

Open maxbrito500 opened 1 year ago

maxbrito500 commented 1 year ago

Feature request

Bark is a Text-To-Audio (TTA) that generates highly realistic voice conversations based on text: https://github.com/suno-ai/bark

It can be customized easily to replicate voices of different genders, ages and even accents thanks to machine learning. The main page provides live examples.

Motivation

There already exist requests for integrating either Whisper or MMS for accurate voice recognition, now we are missing the voice output so that we can have a full Alexa-like system that can work fully offline and privately at our houses.

GPT4All is the ideal platform for combining these features.

Your contribution

Can help with testing. Not sure how to write a plugin for integrating this feature but would be willing to learn.

shiloh92 commented 1 year ago

This would be great!

niansa commented 11 months ago

Hmm, don't think Bark is reliable enough for this task.

Jacobthegr8 commented 1 week ago

Man would that be cool, I could set up a mini PC with an old graphics card with enough V-Ram and 3d print an enclosure to house a computer running this plugin. I just had to sit down and google "How long does spray paint dry" and I KNOW that an LLM would serve that up much quicker than it took me to power on a PC and search for that.

Obviously if possible, I would replace the output voice with GLaDOS for Portal 2.