Is there a good way to have two instances of Cursor running on the same machine (the client running locally and the server running on another computer connected via ssh using the remote dev ext) so that Cursor AI can have both sets of code as context?
Right now I have each set of code update an API doc that I copy to the other when a change is made. This works but the AI doesn’t have access to the other code which limits it’s ability to diagnose.
Cursor will need everything open in the same workspace – we don’t support cross workspace context.
Instead of using remote-ssh, which will require a separate window, could you try using an scp / rsync extension (e.g. Sync-Rsync - Visual Studio Marketplace) that will allow both files to be in the same workspace?
Thanks. I setup a sync-rsync rule on my client that copies each workspace to the other machine one-way, then I put directions in the local rules file to point to the copy. I run rsync when I need to provide context to the other machine. I run two copies of Cursor side by side.
Does this seem reasonable?
I think you should only need the local copy, since that’ll sync/rsync with the remote. But if you need to run commands / use the debugger / etc, … on the remote, then y3es, probably best to have both windows open.
Did Cursor remove the sync-rsync extension? I setup a new system and it’s not there anymore, thus my client-server development is suffering.
After doing some digging, I’m pretty sure the extension isn’t there due to the OpenVSX change.
With all the time I’ve saved using Cursor for coding, now I’ll need to give it back to get my system working again. This is the kind of stuff that gets me looking at alternatives.