WhatsApp AI Reply

Use WhatsApp client library and OpenAI client library to auto-reply to your instant messages.

⚠️ Use at your own risk. The WhatsApp client library is unofficial and could get your blocked from the network.

Usage

Docker

docker run \
  -e SELF_CONTACT_ID=xxxxx@c.us \
  -e OPEN_AI_API_KEY=${OPEN_AI_API_KEY} \
  -e OPEN_AI_MODEL_KEY='gpt-4-1106-preview' \
  -e OPEN_AI_SYSTEM_PROMPT='You are an assistant who expands an emoji response to a message into a full text response' \
  -v ./whatsapp-auth:/mnt/auth-data \
  jmc265/whatsapp-ai-reply

Environment Variables:

SELF_CONTACT_ID - This is the ID that WhatsApp uses to identify you. The value is outputted to the console when you start up the app for the first time.
OPEN_AI_API_KEY - Your API key for OpenAI
OPEN_AI_MODEL_KEY - Choose an OpenAI model
OPEN_AI_SYSTEM_PROMPT - The System prompt to append to each OpenAI request

Volume Mount:

/mnt/auth-data - The location to store the WhatsApp authentication files

Docker Compose

version: "3.4"

services:
  whatsapp-ai-reply:
    image: jmc265/whatsapp-ai-reply
    container_name: whatsapp-ai-reply
    restart: unless-stopped
    volumes:
      - ./whatsapp-auth:/mnt/auth-data
    environment:
      SELF_CONTACT_ID: xxxxx@c.us
      OPEN_AI_API_KEY: ${OPEN_AI_API_KEY}
      OPEN_AI_MODEL_KEY: 'gpt-4-1106-preview'
      OPEN_AI_SYSTEM_PROMPT: 'You are an assistant who expands an emoji response to a message into a full text response'

Local

Make sure you set the environment variables: SELF_CONTACT_ID, OPEN_AI_API_KEY, OPEN_AI_MODEL_KEY, OPEN_AI_SYSTEM_PROMPT and then:

npm install
npm run start