This repository contains a voice-activated chatbot that listens for audio input, detects the wake word (either 'bing' or 'gpt'), and processes the user's request using either the Bing or GPT-3.5-turbo chat model. The resulting response is then synthesized into speech using Amazon Polly, and the audio is played back to the user.
To use this chatbot, you'll need the following libraries:
- openai
- asyncio
- re
- whisper
- boto3
- pydub
- speech_recognition
- EdgeGPT
Please make sure to install all the dependencies using the following command:
pipinstallopenaiasynciorewhisperboto3pydubSpeechRecognitionEdgeGPT
Before using the chatbot, make sure to set up your OpenAI API key and Amazon Polly credentials.
- Store your OpenAI API key in a config.json file:
{"openai_api_key":"your-openai-api-key"}- Configure your Amazon Web Services (AWS) credentials. You can do this by setting up the AWS CLI or by setting the following environment variables:
export AWS_ACCESS_KEY_ID=your-aws-access-key-idexport AWS_SECRET_ACCESS_KEY=your-aws-secret-access-keyexport AWS_REGION=your-aws-region
To run the chatbot, simply execute the following command:
The chatbot will listen for the wake words 'ok bin' or 'ok chat' and respond to your voice commands using the Bing or GPT-3.5-turbo model, respectively.