Hello Cursor community!
I’ve been enjoying using Cursor Pro and the enhanced coding experience it provides. However, I’ve encountered a few challenges regarding the management of API calls and the switching between different models. I’m sharing my setup and the issues I’ve faced, hoping for guidance or feature recommendations.
My Setup
• Cursor Plan: Pro
• Custom API Key: Enabled for gpt-4o to optimize API calls and stay within the plan limits.
• Cursor Composer Models: Occasionally use claude-3-5-sonnet and claude-3.5-haiku for specific tasks.
• Usage Preference: I reserve premium models only for tasks requiring the Cursor Composer’s capabilities.
Challenges
- Custom API Key Disabling:
• I’ve noticed that the Cursor platform sometimes disables the custom OpenAI API key without any explicit action from me.
• This unexpected behavior causes the system to default to the premium models, unintentionally consuming their associated API calls.
- Visibility of API Key Status:
• It’s not always clear whether the custom OpenAI API key is currently enabled, especially when switching between different models or interfaces like Command-K, Chat, and more.
Possible Solutions
I believe addressing these challenges could improve the user experience significantly. Here are some ideas:
- API Key Configuration:
• Add an option to prevent Cursor from automatically disabling the custom OpenAI API key unless the user explicitly disables it.
- Enhanced Status Visibility:
• Include a clear status indicator in the interface (e.g., Command-K or Chat) showing whether the custom OpenAI API key is enabled or not.
I’d love to hear from the community or the Cursor team:
• Are there any existing configurations or workflows to tackle these issues?
• Could these suggestions be viable for future updates?
Looking forward to your thoughts and suggestions!