WAssist

Wassist allows you to contact GPT3 directly from WhatsApp and not only that. Wassist also allows you to save your own personal data and later search and retrieve it using GPT3 to generate a response. In the examples folder, you can see several examples of how to use this bot so you don't have to remember anything ever again.

Supported Features

• No-Command: This is the default mode. The bot will respond to any message you send it through the GPT3 regular completion API.

• Commands: This mode is activated by sending the bot a message starting with the command character. The default command character is /.

  • Currently supported commands:
    • /h: Show a help message that lists all the commands
    • /s : Save the message to the database
    • /q : Ask a question about the database and get a response from GPT3.
    • /f : Find related messages in the database.
    • /w: Get weather forecast from IMS.
    • Send audio file for transcription (Multi-languages).
    • d/ : Draw using Dall-E.
    • /c: Get OpenAI estimated costs
      • Today.
      • Yesterday.
      • Last 7 days.
      • Last 30 days.

Components and Frameworks used in WAssist

• Loguru For logging.
• FastAPI For Whatsapp webhhoks.
• ma-nish Whatsapp Cloud API client.
• weatheril For Weather forecst.
• OpenAI OpenAI client for ChatGPT, Dall-E, and Whisper.

How to use WAssist

Wassist can be installed and run as a system service or a Docker container.

1. open Whatsapp cloud API with Meta and creat an application. read Here how to do it.

2. Set the following environment variables:

   • TOKEN= #Whatsapp API Token
   • PHONE_NUMBER_ID= #Whatsapp phone number ID
   • VERIFY_TOKEN= #Token for whatsapp token verification
   • OPENAI_KEY= #OpenAPI API key
   • ALLOWED_NUMBERS= #List of numbers allowed to communicate with the bot, comma-separated values.

3. If you want to run WAssist as a docker container, copy the following code into your docker-compose.yaml:

   version: "3.6"
   services:
   wassist:
       image: techblog/wassist
       container_name: Wassist
       restart: always
       ports:
       - 80:7020
       environment:
       - TOKEN= #Whatsapp API Token
       - PHONE_NUMBER_ID= #Whatsapp phone number ID
       - VERIFY_TOKEN= #Token for whatsapp token verification
       - OPENAI_KEY= #OpenAPI API key
       - ALLOWED_NUMBERS= #List of numbers allowed to communicate with the bot, comma-separated values.
       volumes:
       - ./wassist:/app/data
   

   **Make sure to set all the environment variables before running the "docker-compose up -d" command.

4. If you want to run WAssist as a systemd service, clone the repository using the following command:

   git clone https://github.com/t0mer/WAssist
   

   enter the Wassist folder and install the dipendencies:

   pip3 install -r requirements.txt
   

   Next, create a file names "wassist.service" under /etc/systemd/system" and paste the following content:

   [Unit]
   Description=GPT Whatsapp 
   After=network-online.target
   Wants=network-online.target systemd-networkd-wait-online.service
   StartLimitIntervalSec=5
   StartLimitBurst=5
   
   [Service]
   EnvironmentFile=/etc/environment
   KillSignal=SIGINT
   WorkingDirectory=/opt/dev/OwnGPT/app/
   Type=simple
   User=root
   ExecStart=/usr/bin/python3 /opt/dev/OwnGPT/app/app.py
   Restart=always
   
   [Install]
   WantedBy=multi-user.target
   

   Make sure to adjust the path for "WorkingDirectory" and "ExecStart" accordingly to the path of the WAssist location

   Next, run the following command to enable and start the service:

   systemctl enable assist. service
   systemctl start wassist.service
   

   To check the status of the service, run the following command:

   systemctl status wassist.servies
   

5. Expose the docker/service ports to the internet so WhatsApp servers can send the messages to your webhooks. this can be done using reverse proxy like Nginx, Traefik or Cloudflare Tunnel. Without this, you will not be able to communicate with the bot.

Acknowledgments

Huge credit and a special thanks to @mangate for creating SelfGPT which my code is based on.