I solved it - the key here is that it’s only in the settings in the Agents Window mode, not the “normal” mode settings. This is confusing!
I’m on 3.9.16 (Mac)
Automations?
I struggled to get this setup. The remote control settings are only available in Agents Window mode, which I’m not using. Also the /remote-control command only appears in Agents Window mode. None of this is documented.
Please can this be ported over to the editor mode? There’s a lot of us not using the Agent Window mode. Editor mode is really the only reason I’m still using Cursor.
Like others have mentioned /remote-control is only available in the Agents window. I only use the Editor window.
But even in the Agents window, using /remote-control (with uncomitted changes) asks if you want to include those changes “in the cloud agent.” What does this mean? I don’t want a cloud agent, I want to control the agent on my local machine.
Are you working on an Android app as my plans for owning an overpriced Apple device are literally never?
We should more clearly document that Remote Control requires the Agent Window, and I’ve opened a PR to update our docs accordingly. There are no immediate plans to port Remote Control over to the Editor Window, as much of its functionality relies on cloud agent-like infrastructure that currently only exists in the Agents Window. That said, we’ll continue monitoring feedback on this!
The “cloud agent” here refers to the remote controlled agent. We should probably find a better term for this in this popup. I’ve shared the feedback with the team!
I tried it in the agent window, and it did nothing. I don’t quite get how it’s meant to work? I type /remote-control and can’t send a message as I have to type something after it - but what? What if I want to be able to control everything on my computer? If I remote control it can I use it on both?
For me the ideal workflow would be that I can just open the cursor app on my phone, see the threads that I have on my computer and continue as normal.
No need to mark one (or many) as remote controlled. It seems unnecessarily complex and requires foresight.
As far as I know until now I’ve not been aware there’s any practical difference between what you can do in the two views. Clearly you want to get rid of the “classic” mode eventually, but I don’t think you’re making this distinction at all clear to users. This is a deeper UX point beyond this specific feature.
Got it. Yeah that would be helpful!
So, I tried using /remote-control and started the agent locally. It doesn’t appear in the mobile app. I’m logged in to the same account, what am I missing?
What I noticed is, if I use /remote-control, it started using cloud agent, so it will cost in different usage pool from Cursor IDE/Chat session?
So I assume it will use API pricing pool instead of Auto/Composer pool. As others also mentioned, I want to control agents on my machine via Mobile app, not starting new cloud agents to pick up works from local agents
I’m running 3.9.16. If I try to type /remote-control in the chat it just tries to make a new skill - it doesn’t recognize that command. I can do agent worker start and then start new chats from my phone to that device, but I can’t access my chats that I start on my devices directly in the cursor app.
Why I must give access to Cursor to my repo on Github to use /remote-control ?
I’m not using Agents Window or any of the cloud agent functionality at all.
I usually work with large workspaces which contain multiple repos, I doubt the cloud agent would work across many repos.
For me I just want a way to control my local transcripts in the editor window from my phone, at the moment I’m using a separate whatsapp plugin for this which I developed.
Yes remote control is enabled
How do you kick off cursor on a repo you alread y are working on your desktop with through this app?
please support ios 17, I still use iphone 13 series. thank you
I’m eager to try the app. Downloaded it, but the first thing I was asked was to change my privacy settings from Privacy Mode (Legacy), which I’m not willing to do (and is apparently non-trivial to revert if changed, according to this HN thread: Tell HN: Installing Cursor on iOS irreversibly changes your privacy settings | Hacker News ).
As @EdLyons noted, it’d be ideal if there was a pure /remote-control feature wherein one could simply continue working with existing agent threads on a local machine in the app without the need to modify privacy settings or provide direct access to GitHub repositories to cloud agents. Perhaps this is something that could be considered for a future version.
I’ve downloaded version 3.9, but I still can’t use the agent function. I’ve enabled remote control in the agent settings, but after typing “/” in the dialog box, “remote-control” doesn’t appear; I can’t start the function at all. I purchased the PRO version; is something missing?