For the past few hours, I haven’t been able to use Cursor smoothly — every task hangs for a long time without making progress, and I have to stop and re-run it each time.
Me too, every task reacts slow, hangs for the first or second actions it takes and then it ends automatically after a long time waiting.
It’s back to normal now on my side.
it is not working again. It is not responding, hangs on … thinking process.
Hey, thanks for the report. Please try the following:
- Run Network Diagnostics: Cursor Settings > Network > Run Diagnostics, then share the results
- Try the
HTTP/1.1workaround: Cursor Settings > Network, switch fromHTTP/2toHTTP/1.1orHTTP/1.0 - Switch DNS to Google/Cloudflare:
System Preferences → Network → Advanced → DNS
Add8.8.8.8or1.1.1.1
Click OK, then restart Cursor
Please share:
- What do the Network Diagnostics show?
- Does switching to
HTTP/1.1help? - Your Cursor version and OS?
This will help us determine if it’s a broader service issue or something specific to your setup.
It is cursor 1.7.54 on mac, it is working but not always on all network modes.
It’s like it successfully replied, modified, provided a solution, and then it will hang on the following reply. I need to stop it manually, resubmit the reply once or twice to make it working again.
Version: 1.7.54
VSCode Version: 1.99.3
Commit: 5c17eb2968a37f66bc6662f48d6356a100b67be0
Date: 2025-10-21T19:07:38.476Z
Electron: 34.5.8
Chromium: 132.0.6834.210
Node.js: 20.19.1
V8: 13.2.152.41-electron.0
OS: Darwin arm64 25.0.0
the most terrible thing is, it will somehow hangs in the middle of a series of long running tasks, it just hang in the middle of django migration
I still have this issue. Restart didn’t help. Temporary I couldn’t open even forum. Now I cannot upload images. Other internet resources work well.
Diagnostics that are failed:
*API
[2025-10-27T06:42:39.093Z] Start
[2025-10-27T06:42:49.103Z] Error: ConnectError: [unknown] Network disconnected
Ping
[2025-10-27T06:42:39.093Z] Sending ping 1
[2025-10-27T06:42:49.105Z] Error: ConnectError: [unknown] Network disconnected
** Chat
[2025-10-27T06:42:39.094Z] Starting streamSSE
[2025-10-27T06:42:49.108Z] Error: Error: Error in the connect call.{"name":"ConnectError","message":"[unknown] Network disconnected","code":2,"metadata":{},"details":[],"cause":{},"stack":"ConnectError: [unknown] Network disconnected\n\tat o.from (/Applications/Cursor.app/Contents/Resources/app/extensions/cursor-always-local/dist/main.js:2:750295)\n\tat /Applications/Cursor.app/Contents/Resources/app/extensions/cursor-always-local/dist/main.js:2:1106131\n\tat async qDt.$callAiConnectTransportProviderStream (file:///Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/api/node/extensionHostProcess.js:161:1585299)"}
* Agent
[2025-10-27T06:42:39.094Z] Starting stream
[2025-10-27T06:42:39.094Z] Pushing first message
[2025-10-27T06:42:49.112Z] Error: Error: Error in the connect call.{"name":"ConnectError","message":"[unknown] Network disconnected","code":2,"metadata":{},"details":[],"cause":{},"stack":"ConnectError: [unknown] Network disconnected\n\tat o.from (/Applications/Cursor.app/Contents/Resources/app/extensions/cursor-always-local/dist/main.js:2:750295)\n\tat /Applications/Cursor.app/Contents/Resources/app/extensions/cursor-always-local/dist/main.js:2:1106131\n\tat async E.stream (/Applications/Cursor.app/Contents/Resources/app/extensions/cursor-always-local/dist/main.js:2:866953)\n\tat async qDt.$callAiConnectTransportProviderStream (file:///Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/api/node/extensionHostProcess.js:161:1585299)"}
Version: 1.7.54 (Universal)
VSCode Version: 1.99.3
Commit: 5c17eb2968a37f66bc6662f48d6356a100b67be0
Date: 2025-10-21T19:07:38.476Z
Electron: 34.5.8
Chromium: 132.0.6834.210
Node.js: 20.19.1
V8: 13.2.152.41-electron.0
OS: Darwin arm64 25.0.0
same proplem here.
Thanks for the diagnostics. The issue is clear: very high ping latency (2179–2772 ms) causing timeouts and hanging.
Interestingly, your HTTP/2 results show everything passing, including Ping.
Since all other checks pass but ping is extremely slow, this suggests a network routing/latency issue rather than a connectivity block. To troubleshoot:
- Test on a different network, try a mobile hotspot to see if the issue persists
- Check for VPN/proxy, temporarily disable if active
- Possible ISP routing issue, if it only happens on your primary network, it may be ISP-specific
Could you share:
- Are you using a VPN or proxy?
- What’s your general location/ISP?
- Does the issue happen on a mobile hotspot or different Wi‑Fi?
This will help determine if it’s a regional routing problem or a local network configuration.
I did try with mobile hotspot and the same issues happened. Not under a VPN or proxy, location is Macau(GMT+8). Didn’t change any ISP or local network settings, it worked perfectly fine in the past few days, the issue only appears since this morning。
oh btw, it was in Auto, agent mode
I have the same problem. For about 4/5 days, AI responses have been extremely slow or blocked. I’ve tried all the support suggestions, but none of them work. It’s impossible to work like this.
Same for me. Yesterday all was good but now only GPT5-high requests worked. All others seem to stuck indefinetly. The same thing is going with WindSurf. Looks like some problems on LLM providers side?
Clarification: model starts to process the request but then loops in endless thinking… Sometimes it just stops in the middle of the work like request is completed but nothing done. SOmetimes work fine. The behaviour is inconsistent.
I use the Auto mode and for me it hangs out every time on the Planning next moves... step.
I turned off my vpn. I tried HTTP/2.0, HTTP/1.1, HTTP/1.0. All Network Diagnostic checks are passed. My location is Russia, Penza. I use wifi internet.
I am facing same issue after waiting lot of time still nothing is working . i have change network still not working
Thanks everyone for the reports. I see that many users started experiencing similar freezes and timeouts today across different regions and network settings.
As I mentioned, we’ve already tried solid troubleshooting steps, but since the issue persists even after changing networks and HTTP modes, it looks like a broader service issue rather than an individual network problem.
For those still affected:
- Which model are you using when the freeze happens? (GPT-4, Claude, Auto Mode, etc.)
- Does switching to a specific model help? (e.g., GPT-5)
- Does enabling a VPN help?
I’ll pass this to the team.
Where does the bug appear (feature/product)?
Cursor IDE
Describe the Bug
When used in Agent or Ask mode chat is taking too much long time to reply. Message was sent at Oct 27, 12:22 PM as of posting this bug Oct 27, 12:37 PM haven’t got reply.
When i cancel/stop the chat event is recorded and billed in the dashboard
Steps to Reproduce
Ask to do task requirement about the project in Agent or Ask mode.
Operating System
MacOS
Current Cursor Version (Menu → About Cursor → Copy)
Version: 1.7.54 (Universal)
VSCode Version: 1.99.3
Commit: 5c17eb2968a37f66bc6662f48d6356a100b67be0
Date: 2025-10-21T19:07:38.476Z
Electron: 34.5.8
Chromium: 132.0.6834.210
Node.js: 20.19.1
V8: 13.2.152.41-electron.0
OS: Darwin arm64 23.6.0
Does this stop you from using Cursor
Yes - Cursor is unusable
Same here, check: Reddit - The heart of the internet
Same here. After the last update cursor is unusable due the frequent freezing, suuuuper slow responses.
very often it gets stuck on the “planning next moves”






