WhatsApp GPT is a small open-source experiment by Daniel Gross that connects WhatsApp to ChatGPT, letting you hold AI conversations inside your WhatsApp chats. Built in the first wave of ChatGPT enthusiasm, it bridges two worlds that had no official integration: a Go program runs your WhatsApp account while a Python process drives ChatGPT in a browser window, and the two relay messages back and forth. The repository has about 3,100 stars on GitHub.
The architecture is charmingly pragmatic. The Go side uses a WhatsApp client library to send and receive messages from a dedicated phone number; the Python side (server.py) controls a real ChatGPT web session in a separate browser window, typing incoming WhatsApp messages into ChatGPT and returning its replies. A bonus script, multichat.py, removes the human entirely and lets two ChatGPT instances talk to each other. The author is candid that the README is sparse and first runs rarely work without tinkering — this is a hacker’s proof of concept from 2023–2024 rather than a maintained product, and it depends on unofficial interfaces on both sides. Even so, it remains a frequently referenced blueprint for anyone prototyping messaging-app bots or studying how early builders wired ChatGPT into everyday tools.
WhatsApp GPT is completely free and open source under the MIT license. You provide the phone number, the computer to run it on and your own ChatGPT access.
WhatsApp GPT is for tinkerers and developers who want to experiment with connecting messaging apps to ChatGPT and are comfortable debugging an unofficial, community-built integration.