I’m on a team plan and hosted a local llm in lm studio via ngrok. I’m using a windows machine. I keep getting the below error when trying to send a prompt:
Request higher limits to continue using Cursor
Can I not do this when on a team plan? Why is it looking if limit is reached for a local llm? I have turned off all other llms in Cursor.
Hey @quldude, thanks for the report!
Your LM Studio setup is reaching Cursor fine, and Cursor recognizes your model as a custom API key model. “Request higher limits to continue using Cursor” is a usage limit block, not a connection issue.
On Teams plans, requests to models you run with your own key (including local ones) still carry the Cursor Token Rate of $0.25 per million tokens, so they count toward your usage limits. Your included usage for this cycle is spent, and your team has on-demand usage turned off, so those requests are stopped before reaching your ngrok endpoint. Auto still works because your team has a separate allowance covering only Auto.
To unblock: a team admin needs to enable on-demand usage (set a team spend limit) in the Cursor dashboard under Settings. You can also click “Request Limit Increase” on that dialog to send them the request. Once enabled, only the Cursor Token Rate is billed, not model cost. Otherwise, it resumes when your usage resets.
One more thing to check afterward: the one request that did reach your LM Studio endpoint came back completely empty (no text, no finish reason). Make sure the loaded model supports tool calling and streaming.