Hi @knaig
Did it work before? Or has it been like this from the start?
Hi @deanrie.
Sorry. Slight change here.
It worked perfectly until 2 days ago. Yesterday, it gave a message for o1 “Vision rate limit hit”. So I decided to give it a break. Today, it is freezing with all models as well with the “generation” message. There is no rate limit message today.
1 Like
Have there been any changes in your network in the last couple of days? Also, did you install any additional extensions? Does the same thing happen in other projects? If yes, try loading without extensions and see if that fixes the problem. You can load from the terminal with this command: cursor --disable-extensions. If the problem persists, try reinstalling Cursor using this guide:
Also, check if these addresses are accessible:
Hi @deanrie, Thanks for the quick responses.
An additional finding. The chat section seems to work fine in general (have not tried with o1 model yet). Its the composer tab which freezes (for any model) .
I tried reinstalling and then running with extensions disabled. same finding.
All the apis, except the 2 below, are accessible from the browser.
https://cursor-cdn.com/ → Error 404 (object not found)
https://download.todesktop.com/ → Bad request
Could you also check if you’ve hit the request limit?
I did. Not hit the request limit as yet on any of the models.
Please provide the following information:
Cursor composer tab freeze issue:
-
image attached.
-
type query in compose. Press enter.
-
Version: 0.42.3
VSCode Version: 1.93.1
Commit: 949de58bd3d85d530972cac2dffc4feb9eee1e40
Date: 2024-10-16T17:56:07.754Z
Electron: 30.4.0
Chromium: 124.0.6367.243
Node.js: 20.15.1
V8: 12.4.254.20-electron.0
OS: Windows_NT x64 10.0.22631
-
No VPN or Zscaler.
-
Error found in console. Attached. Looks like cursor is trying to attach a non-existent image file. I am not trying to attach any file myself.
Try clearing the cache, and in this case, replace VSCode with Cursor.
Thanks. Its working now after reinstallation and deleting all the histories, but composer is asking me for the context. Earlier, it was taking the context by default. I have several files, so I want to just provide the root directory as context. How do I do that?
You can add context using @.
Same thing on my side. I think it started freezing a couple days ago.
The composer would start generating a file and it will freeze midway, I haven’t ran out of fast requests.
Perhaps when you start generating you need to do it in multiple steps and the first goes through but then something breaks on the following? When generating with o1-mini it successfully generated my code, but then froze on the Applying step (I guess this is making the diff). So it seems it freezes regardless of the fact if the next thing is another generation prompt or it is follow up step in the processing you do for result - it will freeze.
I waited for a day, then wiped and reinstalled Windows. I was facing lots of issues. The reinstall seems to have fixed all of them. Rate limit though might have been unrelated to the OS itself.
I found a workaround for this that I often use in ChatGPT: simply start a new chat to stop the freezes.
This issue often occurs when you have a long chat history. It seems that the AI needs to quickly process the entire conversation each time before responding to a new prompt. When the history gets too long, it takes longer for the model to process, and the more advanced the model, the longer it takes—sometimes causing it to time out.
If this works similarly to ChatGPT, there may come a point when opening a new chat no longer fixes the problem. In that case, you would need to delete older chats to free up memory.
Hope this fixes it for anyone having the same issue.