I’m experiencing an issue with invoking an Action from the GPT, specifically when using OAuth authentication and subsequent API calls.
The setup is intended to work as follows: A user enters a prompt, which triggers a sign-in process via OAuth. After a successful sign-in, confirmed by a “You have successfully signed in via OAuth” message, the system is supposed to automatically invoke the first in a series of predefined Actions (API endpoints).
However, the issue I’m facing is that after the successful OAuth sign-in, the first Action is not being triggered as expected. The UI shows an indication (via the down arrow) that information is ready to be sent to the API, but there’s no actual call to the endpoint or any error message displayed.
I’ve attempted several troubleshooting steps such as deleting and re-adding the Action, and toggling the authentication settings.
Interestingly when I disable the authentication and send the same prompt, it prompts asking to ‘Allow’, ‘Always Allow’, or ‘Deny’ access. When I select ‘Allow’ I receive a failure message in ChatGPT and a corresponding HTTP 401 error in the backend log as expected. Yet, when authentication is enabled, there seems to be no activity or error logging in either the front end or back end.
Login success message:
API call that is supposed to be invoked after login (But not triggered)