I am a user with 500 request per month. Recently I found my Cursor often stop running with no reason, usually after calling some tools. So I have to spend another request to ask Cursor to “continue”, which will cosume a lot more “requests” for one task.
I am wondering if this is a bug or network problem, or just Cursor maks it on purpose to cosume more “requests”?
BTW, my Cursor often " reconnecting…", though it will usually go back to work after connecting.
Hey, those symptoms look familiar, there are a few similar reports:
- Cursor keeps interrupting its responses and can't execute completely at one time
- Cursor Stalls After Exploring Files — Forcing Me to Manually “Continue” (Worsening in Recent Weeks)
- Request stuck at "Reconnecting" and still consuming requests
Usually it’s a combo of two things:
- Agent stops after tool calls. In some cases this is a bidi stream bug where messages arrive out of order and it hits
DeadlineExceeded. We’re tracking it, no ETA yet. - Reconnecting… This is most often network instability to our backend. In our logs your requests arrive without server errors, so it looks like the connection is dropping on the network side. For users in China this is pretty common due to latency and packet loss to US infra.
To narrow down what’s happening for you, please share:
- Cursor version Help > About and your OS
- Request ID from the last broken session right upper corner of the chat > Copy Request ID
- Run Cursor Settings > Network > Run Diagnostics and paste the result
On billing, each continue is a new request, so that behavior is expected. The real issue is the connection dropping. Temporary workaround, try a stable VPN, it often helps with reconnecting.