2.2.44 works fine… so far 
In 2.3.9 I get these errors on any request to model. And it still not usable at all.
Request ID: 6cf48b75-d3f5-4efd-9079-1b3a1bfb5ae8
{“error”:“ERROR_CUSTOM”,“details”:{“title”:“Agent Execution Timed Out”,“detail”:“The agent execution provider did not respond within 4 seconds. This may indicate the extension host is not running or is unresponsive.”,“isRetryable”:false,“shouldShowImmediateError”:true,“additionalInfo”:{},“buttons”:[{“label”:“Reload Window”,“reloadWindow”:{}}],“planChoices”:}}
network test is OK
my experience after following this thread - 1) had to do shared (no privacy) , deleted indexes, EVERY request has to be a + new request (so I lose context every follow up), I can at least get 1 command through before it times out.
I am having the same issues: (first one is 2.3.14 version, the second is 2.2.44)
I’m about to give up and look for a Cursor alternative. I ‘m stuck. It stuck on “Planning next actions”
It’s truly shocking that this issue still hasn’t been fixed. After repeatedly suffering from this problem, I rolled back to version 2.0.77 and have been too afraid to upgrade since then. Today, checking on this topic, it turns out it’s still not resolved! The Cursor team should halt the development of other new features and focus their efforts on solving this issue—it really might drive many people to switch to other platforms, and I’m starting to waver myself.
3 Likes
I’m also speechless with the cursor development team; it took me over half a day to figure out that I had to use version 2.077.
I too encountered this issue in version 22.4 and was extremely disappointed that it remained unresolved for so long.
It’s unbelievable that they keep reintroducing this bug. So sick of it.
1 Like
Great. Reward them for failure by giving them more money!
I ran into this today. For me it turned out that my git program was not functioning properly. Once I got that sorted out the problem went away. Just something you might want to try.
DON”T WORK
Version: 2.3.29
VSCode Version: 1.105.1
Commit: 4ca9b38c6c97d4243bf0c61e51426667cb964bd0
Date: 2026-01-08T00:34:49.798Z
Electron: 37.7.0
Chromium: 138.0.7204.251
Node.js: 22.20.0
V8: 13.8.258.32-electron.0
OS: Darwin arm64 25.2.0
The agent execution provider did not respond within 4 seconds. This may indicate the extension host is not running or is unresponsive.
Today, after update to Version: 2.3.29
Commit: 4ca9b38c6c97d4243bf0c61e51426667cb964bd0
Date: 2026-01-08T00:34:49.798Z
Browser: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Cursor/2.3.29 Chrome/138.0.7204.251 Electron/37.7.0 Safari/537.36
Request ID: 4ea8d23a-ae06-4d1f-a06b-b65f433bfbe7
{“error”:“ERROR_CUSTOM”,“details”:{“title”:“Agent Execution Timed Out”,“detail”:“The agent execution provider did not respond within 4 seconds. This may indicate the extension host is not running or is unresponsive.”,“isRetryable”:false,“shouldShowImmediateError”:true,“additionalInfo”:{},“buttons”:[{“label”:“Reload Window”,“reloadWindow”:{}}],“planChoices”:}}
Im dissapointed that this bug is constantly coming back. Fcuck IT
1 Like
I ran into it again and again and again. Burning my budget for nonsense!
Case 1: Command execution stuck
Case 2: Cannot generate code changes:
（I manually terminate it after waiting for 3min.)
I experienced these f**king errors about once every three editings, and I dare not leave my computer. If I have to constantly monitor the Agent, manually edit it, stop it, and roll it back, then what’s the point of using Cursor??? I will opt for other products
Look at this screenshot. Seriously, look at it.
“String to replace was not found.”
“String to replace was not found.”
“Incorrect tool arguments.”
“Cancelled.”
I’m an annual subscriber with 4 months left (approx. $80 value), but using Cursor is actively costing me MORE money because your agent is too incompetent to perform a simple find-and-replace or run a basic npm command without hallucinating.
I’ve wasted $10 worth of usage this month just watching your AI fail in loops. The way you count failed, hallucinated operations against user quota is absolute **theft**. At this point, writing off my remaining subscription feels cheaper than continuing to pay for this broken error management.
Fix your agent’s context awareness or stop charging for failed tool calls. This is ridiculous.
Sometimes I needed to wait 1-2 minutes for the Cursor to respond.
However, at other times, I left the chat for 30 minutes, and it was still marked as “Planning next moves”.
This issue has persisted for a long time and remains unresolved.
You might be able to chat now. But in 5 minutes, that’s not guaranteed.
UPDATE:
I have a feeling that Cursor is queuing user requests.
Because “Planning next moves” takes a very long time. But no request is being made to the provider.
I waited 5 minutes, and one request reached the provider. Immediately after that, Cursor responded.
It says “Asking questions” but I do not see any questions. (
This lasts for 10 minutes, then I stopped it.
I think this could also be part of the reason.
When I enable a custom OpenAI endpoint, the Cursor becomes unstable.
With the same settings, sometimes it works perfectly, but sometimes it gets stuck at “Planning next moves”.
I checked the console and got the error:
We encountered an issue when using your API key: Provider was unable to process your request\n\nAPI Error:\n\n\nError\n
Although the endpoint is still working fine.
Could there be a queued request or a firewall on the server side?
I tried:
api.openai.com => works fine
any other domain => failed.
Note that it doesn’t fail all the time. It’s random.
