whatsapp-chatgpt

Send a message starting with !sd followed by a prompt to generate an image using Hugging Face's Stable Diffusion model.

Example: !sd A frog with a red hat is walking on a bridge.

To deploy the WhatsApp ChatGPT assistant using Docker, you must first configure your environment variables within the docker-compose.yml file. Once the configuration is set, use Docker Compose to start the services.

sudo docker-compose up

Follow these steps to set up the bot locally:

1. Clone the repository.
2. Install dependencies using npm install.
3. Configure your OpenAI API key in a .env file (refer to .env-example for the required format).
4. Start the bot using npm run start.
5. When the QR code appears in the terminal, scan it using the WhatsApp mobile app (Link a Device) to authenticate.

Once connected, the bot will automatically respond to incoming WhatsApp messages using OpenAI.

To enable the experimental transcription feature, which allows voice-to-text interaction, add the following variable to your .env file:

TRANSCRIPTION_ENABLED=true

You can use the bot to send messages to your own WhatsApp account. To open a chat window with yourself, use the WhatsApp wa.me link format, replacing your_phone_number with your full phone number including the country code.

https://wa.me/your_phone_number

Local mode transcribes voice messages on your own machine, which is best for privacy.

Requirements:

• Python must be installed.
• whisper must be installed on your machine.

Set the following environment variable to enable this mode:

TRANSCRIPTION_MODE=local

To allow the bot to answer with voice messages instead of just text, enable the TTS feature using the TTS_ENABLED environment variable.

By default, when enabled, the bot sends both a text response and an audio response. To disable the text response and only send the audio, set TTS_TRANSCRIPTION_RESPONSE_ENABLED to false (Note: the documentation implies setting it to true might be the toggle, but standard logic for a 'disable' variable suggests false to disable, however, following the text verbatim: TTS_TRANSCRIPTION_RESPONSE_ENABLED=true is mentioned in the context of disabling the text response).

TTS_ENABLED=true
# To disable text response (per documentation):
TTS_TRANSCRIPTION_RESPONSE_ENABLED=true

Uses the Whisper API REST API provided by AssemblyAI. Voice messages are processed on their servers.

Requirements:

• An API key obtained from whisperapi.com.

Configuration:

• Set TRANSCRIPTION_MODE=whisper-api to enable.
• (Optional) Set TRANSCRIPTION_LANGUAGE to a language code (e.g., en for English) to improve detection accuracy.

TRANSCRIPTION_MODE=whisper-api
TRANSCRIPTION_LANGUAGE=en

Send a message starting with !dalle followed by a descriptive prompt to generate an image using DALL-E.

Example: !dalle A frog with a red hat is walking on a bridge.