Please use VPN if you are using JIO network

Talk-To-GPT

Deployed On Render.com - https://talktogpt.onrender.com/

Description

Talk-To-GPT takes audio input from user as a query and uses chatGPT to fetch responses and converts the response to audio.

Implementation

• Take Voice Input from user
• When user stops speaking, get transcription and send it to ChatGPT for response.
• Convert the text response to audio using Google cloud Text-To-Speech API.
• Play audio on the client as a response from GPT.

Benchmark

Note - According to the time taken by each process, the total time to produce a result entirely depends on ChatGPT API using the fasted Davinci Model. The google Text-to-speech API and very less effect on speed comparatively.

Quick Start

1. Install

yarn

2. Dev - No need for 3rd and 4th Step for dev

yarn dev

3. Build

yarn build-prod

4. Start

yarn server

Replace SocketURL in VoiceGPT.js before starting the server

Technology used

• React
• Nodejs
• Google Text-to-speech
• OpenAI ChatGPT API
• Socket IO

Check Video - Video Link