Skip to content

WeberJulian/AI-voice-chat

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

31 Commits
 
 
 
 
 
 
 
 

Repository files navigation

AI Voice Chat

Overview

This is a simple react app that allows you to chat with an AI assistant using your voice.

It uses Whisper large v3 for transcription, openchat 3.5 AWQ for the AI assistant, and XTTS v2 for text-to-speech.

Its main feature is speech-to-speech latency, it more than halves the latency of the ChatGPT with voice demo video. This repo runs on a single RTX 3090 GPU.

No concurrency is supported and the project is not optimized production ready. It's also probably riddled with bugs so if you experience some, please open an issue or send a PR.

The XTTS v2 model is from the coqui-TTS. If you have any questions about the model or the project, you can join our discord server

Demo

ai_voice_chat_demo_3.mp4

Installation

Requisites

  1. Have a Nvidia GPU with more than 16GB of VRAM and latest drivers
  2. Have docker-compose installed

Steps

  1. Clone the repo
git clone https://github.com/WeberJulian/AI-voice-chat.git
  1. Build the react app
cd AI-voice-chat
cd web-app
npm install && npm run build
  1. Start everything 🚀
cd ..
docker-compose up

Usage

  1. Open the app in your browser at http://localhost:5000

  2. Allow microphone access

  3. Push to talk either with the Shift ⇧ key or the circle

  4. Enjoy!

To reset the conversation, refresh the page.

Custom models

If you fine-tune XTTS and want to use your own model, you can add that line to the docker-compose.yml file, in the tts service:

services:
    ...
    tts:
        ...
        volumes:
            - /path/to/your/model:/app/tts_models

In the /path/to/your/model folder, you must have the following files:

  • config.json
  • model.pth
  • vocab.json

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published