best practices on using my own Azure API key for Cursor? I got it to work by deploying a gpt-4o model in Azure’s AI studio, and then using that API key, but then the way it works with Cursor UI is a bit confusing. Is that key then only always going to call gpt-4o, even if I switch the model I’m using in Cursor? I assume yes, and if that’s the case, is there another configuration method to have access to multiple models within cursor while using Azure deployments?
I don’t think so. Once the API key is on, all your connections go through it. You can also manually add any models you need that aren’t in the list. When you switch models, if they’re available in Azure, everything will work fine, and you’ll get responses from them. If you need models from the Cursor subscription, you’ll have to turn off your API key so Cursor knows how to handle requests.