I’ve been experiencing an issue where my Cursor Agent keeps getting stuck in the “Generating” state indefinitely. It doesn’t seem to be a billing issue — I’ve verified that my OpenAI/Gemini accounts have sufficient quota.
Here’s a summary of what’s happening:
The Agent appears to load correctly, but once it starts generating, it stays stuck on “Generating” and never completes.
I’ve tried switching models (GPT-4, Claude, Gemini), restarting the Agent, and even restarting Cursor itself — nothing helps.
Network connectivity is stable, and my VPN setup is working fine (I can access other AI services normally).
There’s no error message, just an infinite “Generating…” loop.
I’ve also verified that the prompt and the task logic are not overly complex or recursive.
Has anyone encountered the same issue and found a fix?
Any insights on whether this is a known bug or something I can resolve from my side would be appreciated.
I wont be much help but when this has happened to me and it doesnt happen a lot, I generally press stop, then start again, after a few attempts it works.