I currently have my chatbot accessible through Cloudflare tunnels with authentication. On a network that blocks *.openai.com, when I enter my API key, it gets blocked when validating. I assume adding it to the config file would bypass the initial check... at that point, does the server hosting Chatbot handle all communications with openai.com or is the client browser still making calls to openai?

Ah, yeah, i understand that for sure. Let me rephrase though (meant it as a general question):

Other than validating the API key, does the server running the interface act as a proxy for the API or are the API calls still made client side? I am running the interface on a home server and accessing the server from a network that has openai.com blocked. My hope was that all requests/responses to/from api.openai.com would be handled server side. If that isn't how this is made, that's fine, just trying to make sure I understand it.

Thank you.

@wescrockett This is a client-side application, so all requests will be executed client-side. Unfortunately, In your case that will mean that requests will be blocked.

This is a client-side application, so all requests will be executed client-side. Unfortunately, In your case that will mean that requests will be blocked.

Thank you for the clarification.

has this been merged? Alternative end-point to openai would be very helpful to have.

has this been merged? Alternative end-point to openai would be very helpful to have.

I don't think this PR is required anymore. It looks like my changes were copied and merged in e60272e