Same issue with me too. I tried @ellengogogogo suggestion the disable http2 but no success. Anyone else?
did nothing for me. Still getting the error.
For me opening a new chat helped. Yes i tried restarting Cursor but that fixed it only for 10minutes.
All, I tried the recommendation to start a new chat window and that indeed fixed the communication problem .. and created a new problem .. where exactly am I?
This is a serious issue for @cursoradmin to fix. I was right in the middle of a significant change when cursor went AWOL. Now after about 5 days, opening a new chat window re-establishes communication, but at what point. I am guessing I have at least a day of reorganizing my thoughts and figuring out where exactly I am. This is a great tool, and I know there are going to be problems .. but this is a big one. Please take whatever steps are necessary to a) document progress so I can recover from another communication failure, b) fix the communication failure, and c) do a MUCH better job communicating with us.
I really like this product, but outages like this are really, really bad.
- John
same here did nothing
Try deleting or renaming the cursor dir and factory reset cursor. On a mac the dir is in the library/application support/cursor.
1 Like
I’m using Mac OS. I removed the cursor with AppCleaner, installed again, and then it worked. I mean I did the clean install. I guess some of the cached states in files make this abnormal behaviour. What log can you help to debug? Please guide me for the debug logging.
Same issues 8e023277-36a9-491b-b95b-8ed69279cc38
Before the update it was a LOT smoother.
A sizable week-long outage with no feedback or reasonable at all is mad for a paid, closed-source software. I’m sure many of us would love to help debug as this is an otherwise great product, but frankly this and the python env errors in a similar state + the difficulty to upgrade versions (“check for updates” and “attempt update” fail pretty much all the time) have sadly made me move on to a competitor.
If it’s of any help, it seems like having a VPN installed (even if not turned on) seems to be a common denominator sometimes but it’s difficult to pinpoint exactly what’s happening here
1 Like
I do have OpenVPN installed, but not running
Hey all,
As many users will know from the past, we’ve faced issues with getting enough compute available for everyone wanting to use the Anthropic models.
Between Anthropic’s own API, and other providers of the Claude models, we are maxing out our capacity across the board.
This has unfortunately meant that as many of you have seen sometimes queries can fail during peak times. While we eventually got this under control for Claude 3.5, the release of 3.7 exacerbated the issue once again.
A week ago, this was in a poor place, but our team has been working hard to resolve this behind the scenes.
As such, we are hopeful that we are in a much better place now regarding connection errors due to model availability. All paid users who are not using slow requests should see little to no connection errors moving forward.
For anyone seeing connection error still, please try the following:
-
If Cursor has never worked for you, or he’s not working in a new network environment, it may be that your network does not support HTTP/2. We now have a workaround for this, which can be enabled by turning on the ‘Disable HTTP/2’ setting. More info on this can be found here: Cursor – Common Issues
-
If you are still seeing intermittent failed requests and have a specific model selected, I’d recommend enabling the automatic model setting we have some intelligent rooting behind the scenes, that should mean you see a lot less errors, well always having a premium model respond to your query!
-
If you see errors repeatedly, in the same project or across all of your projects, and you are totally unable to get a response from Cursor, please let us know, as we’d love to diagnose what is going on there. You may be able to get things working by following our troubleshooting steps, but please let us know in this thread what issues you are seeing, if the above doesn’t help.
Her sis my setup, no connection at all now after update.
even with my own Anthropic API being used makes no difference, …surely I should be able to use my own model via api at least if the connection to Curson fails
Version: 0.47.8
VSCode Version: 1.96.2
Commit: 82ef0f61c01d079d1b7e5ab04d88499d5af500e0
Date: 2025-03-18T06:55:51.040Z (1 day ago)
Electron: 32.2.6
Chromium: 128.0.6613.186
Node.js: 20.18.1
V8: 12.8.374.38-electron.0
OS: Darwin arm64 24.1.0
Version: 0.47.8
VSCode Version: 1.96.2
Commit: 82ef0f61c01d079d1b7e5ab04d88499d5af500e0
Date: 2025-03-18T06:55:51.040Z (1 day ago)
Electron: 32.2.6
Chromium: 128.0.6613.186
Node.js: 20.18.1
V8: 12.8.374.38-electron.0
OS: Darwin arm64 24.0.0
Same issue
With the latest updates, all my issues are fixed now. I didn’t change anything other than installing the updates. Although I didn’t need to do this but there are instructions now by Cursor team for some optimization changes they have done and that may not work in all environments. Specifically regarding HTTP/2(@danperks has already posted this above) Cursor – Common Issues
I’m also experiencing the same issue, and it’s getting quite frustrating. Honestly, with these ongoing problems, I’m starting to wonder if this tool is really worth it.
20a6b88b-5575-4e33-b251-ea80057e3362
getting plagued with it this morning
Hey, can’t see that connection ID internally, can you try the following:
一只在等待或者提示VPN错误。
Faced the same problem. Tried all the tips from the Common issues tab, nothing helped. I completely deleted the cursor via App cleaner, the tab cursor never worked, ai chat worked once and then also gave a connection error.
Request ID: fc95c664-70b2-4b73-8193-de473672bca6
upd. I switched http1 and htpp2 several times and ai chat worked, but the tab cursor still doesn’t work
This is a caching issue. Same thing happened to me and I couldn’t connect. Removing .cursor file should work or you should create/open a new project. This was what worked for me.