WhatsApp is a trademark of Meta, inc.
WhatsGPT is a Proof-of-Concept project to build a WhatsApp chat bot using official WhatsApp API. The bot would be powered by OpenAI GPT-3 to interpret messages sent to it, and fulfil it against realtime data that Backend module has access to.
Main component of the project is Backend. To start it, you can choose one of the following three options:
- Start Backend from python directly: Follow Contributing Guidelines to create Python
venvand installrequirementsonly (it is safe to skipdev_requirements), set working directory to Backend, and start it usingpython .. - Run docker image*: Set working directory to Backend, build image using
docker build -t whatsgpt-backend ., and run it withdocker run --rm -it whatsgpt-backend. - Run docker-compose*: Set working directory to project root, and run docker-compose as
docker-compose up. When done stop it withdocker-compose down. If you have pulled updates to Backend, or you made your own, run it asdocker-compose up --buildto force rebuilding image of Backend.
* If you are using podman, you can follow the same steps as docker, docker-compose.
Following serves as quick overview to project progress:
- WhatsApp API Webhook:
- User should be able to send message to chat bot.
- User should receive an echo to message he sent.
- GPT-3 Integration:
- Rather than getting echo of the message, User should receive an answer as provided by GPT-3.
- Realtime Data Integration:
- Rather than getting responses compiled by GPT-3, User should receive an answer that reflect information bound to realtime data.
- Messages Log:
- Admin should be able to view a log of all messages and responses from Dashboard.
- Admin should be able to mark certain messages with different markers for later action.