Auto, Sonnet, GPT, any of them - service is not stable
For AI issues: add Request ID with privacy disabled
Request ID: 48c5f093-97b6-4412-b8cb-4ca226b2a6c2
{“error”:“ERROR_OPENAI”,“details”:{“title”:“Unable to reach the model provider”,“detail”:“We’re having trouble connecting to the model provider. This might be temporary - please try again in a moment.”,“additionalInfo”:{},“buttons”:,“planChoices”:},“isExpected”:false}
ConnectError: [unavailable] Error
at Iqc.$endAiConnectTransportReportError (vscode-file://vscode-app/c:/Program%20Files/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:7337:399337)
at nwo._doInvokeHandler (vscode-file://vscode-app/c:/Program%20Files/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:490:36007)
at nwo._invokeHandler (vscode-file://vscode-app/c:/Program%20Files/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:490:35749)
at nwo._receiveRequest (vscode-file://vscode-app/c:/Program%20Files/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:490:34514)
at nwo._receiveOneMessage (vscode-file://vscode-app/c:/Program%20Files/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:490:33336)
at WDt.value (vscode-file://vscode-app/c:/Program%20Files/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:490:31429)
at _e._deliver (vscode-file://vscode-app/c:/Program%20Files/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:49:2962)
at _e.fire (vscode-file://vscode-app/c:/Program%20Files/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:49:3283)
at agt.fire (vscode-file://vscode-app/c:/Program%20Files/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:7322:12156)
at MessagePort. (vscode-file://vscode-app/c:/Program%20Files/cursor/resources/app/out/vs/workbench/workbench.desktop.main.js:9845:18403)
Additional Information
really tried of having to debug your application every time we want to use it.
Hey, Cursor had some issues in past few days but it’s definitely not a common occurrence, it should be working fine at the moment.
Do you use any VPN or maybe specific DNS settings?
This specific issue seems to be more OpenAI related, are you sure it still happens with other models?
Try again in next few hours and feel free to tell me how it went!
Hello. @Ra.in - i would really ask to go “beyond your standard script” - see below and i would like to see something a bit more deep than “it must be something on my end” since i believe i have shown clearly that its an issue in your applications.
no - not using VPN, direct connection - no changes.A
What i have just seen is that following the update yesterday if i try to use /continue with a context I get the error.
if i start a new windfow with new profile then magically your chat “works”..
Sadly, it means i have to redo all the contexts from yesterday whcih, including your issues with Char Set injections causing each prompt to run 2 or 3 times means that i have to repeat everything done this week.
Please stop trying to push this off on local network connectivity or client issues and explain why it is that a context that “worked” yesterday, when trying to continue ttoday throw a “Model unavaiable” error - and yet, if we throw away our work and start fresh, the your model connectivity works?
Here are two side by side screen shots - so can you please exaplin to me why/where you would star twith “its is a VPN issue???” - its so disrepectful to try to allay a clear product fault on the end customers equipment..
grrrr! - you guys need better processes!
so - given the left if a ‘new window” and it works and the right window is a continuation of the work from yesterday that fails - how and where it would be a “local connectivity issue” on my side.
Since the functions work in a new window, when / how long should i ‘wait” to continue with the work from yesterday?
Based on your screenshots showing new windows work but old ones don’t, it does look like those specific chat contexts got corrupted somehow rather than a broader outage or network issue. Your first message didn’t mention this distinction - it sounded like a complete service outage, which is why Ra.in started with standard network troubleshooting. Have you tried duplicating the affected chat to see if that preserves your context while bypassing whatever corruption occurred? That might save you from having to redo everything.
This happened to me today as well. I was able to fix it by logging out of Cursor and then logging back in. Go to Cursor Settings → General tab → Log Out (very bottom). Then just log in again. Hope this helps you as well.
Im sorry - statements like “it looks like this or that got corrupted” following the installation of an update is not really a adequate response at all.
All day with your update, your new update refused to read and see the existing chats, when i try to export and then import (or even paste back in to summarize) the application just ‘loops” and doesnt take the paste.
In fact, anything with my context and working chat for the last two days no longer works at all, even a simple prompt where the agent comes back to say “ I will go ahead and do this” starts to plan and then instantly drops out, requiring me to ‘reprompt”, and the in that reprompt, it starts “from scratch” again.
So that i understand, we should not ever expect that from one update to another that anything we did before should carry over and really not depend on being able to continue work from one day to the next in the same context or chat, rather we should expect that every time we prompt is a “fresh one” and things like using the context cannot be relied on?
spent 7 hours to day trying to get back to where I was at 5pm yesterday -= your own errors show a “Transient Network issue” when in fact it has nothing to do with networking or model availability at all resulting in more time lost.
I dunno - seems pretty shoddy if each time is a new context, and we cant use / build contexts over time that can persist beyond a day or even import an exported chat context - basically its back to copy paste from GPT (singular contexts) huh?
not really impressed today at all and id love to you show you things, but the agent cannot even execute actions now without dropping out requiring reprompting.
And yet another update sent today - so should i risk downloaded it and wiping out my context from today (which i mean, is nothing, since its just me spending the day trying to understand why the agent that worked Tuesday doesnt now)..
