--- title: "Cursor Remote Control: Control Local Agents from Your iPhone" description: "Cursor Remote Control (Oct 6) lets the iOS app reply to local agents. Pairing steps, enterprise toggle, Projects vs. Environments API, and network checks." url: https://cdnping.org/en/blog/cursor-remote-control-projects-environments language: en published: 2026-10-18 category: "Developer tools" tags: ["Cursor","Remote Control","Cloud agents","Projects","Environments API"] source: cdnping.org license: Free to quote and summarize with a link to the url above. --- # Cursor Remote Control: Control Local Agents from Your iPhone You can now control a Cursor agent from your phone. On October 6, 2026, Cursor shipped Remote Control: once you pair the iOS app with the desktop app, you can view and reply to local agents running on your computer. The agent still runs on your machine, and the phone is just a remote. Cursor also launched Projects (beta) on September 10, and on October 16 added shareable canvases for cloud agents and put the Environments API into beta. Everything below follows Cursor's official changelog, as of October 2026. ## What Remote Control is, and what it isn't According to the [Cursor changelog (October 6)](https://cursor.com/changelog/remote-control-local-agents): - From the iOS app you can see what a local agent is doing and send it messages. - The agent does not move anywhere. It keeps running on your computer, and the app connects to it. - It does not depend on cloud agents. - It is iOS only for now. The official page does not mention Android. So it solves the "I stepped away and the agent is stuck waiting for approval or one more instruction" problem. It is not a way to hand the task off to the cloud. ## How to pair 1. Download and sign in to the Cursor iOS app on your iOS device. Use the same account as on the desktop so your computer shows up in the list (follow the in-app prompts if they differ). 2. Your computer appears in the app automatically. Tap it. 3. Back in the desktop Cursor, approve the pairing request that pops up. 4. After pairing, your local agents appear as a list, where you can watch activity and send messages. One more thing before you leave: the computer has to stay on and online. If you don't want it to sleep, turn on "Keep this computer awake" under Remote Control in the desktop settings. Per the official docs, to prevent sleep while you're away the computer must be plugged in and the lid kept open. A laptop closed and tucked into a bag does not meet that condition. ## Enabling it for enterprise The official note says the feature is on by default for everyone except Enterprise organizations. For Enterprise, an admin has to enable it under Org settings > Security and identity > Remote control. Whether to allow it depends on your company's security requirements for remote access to developer machines. ## Projects vs. Remote Control [Projects (beta, September 10)](https://cursor.com/changelog/projects) is aimed at bigger pieces of work, such as a feature, a migration, or a whole app: - A coordinator agent plans and delegates. It doesn't write code itself, and it can run multiple implementation agents in parallel. - It runs on a separate cloud computer, so it keeps going even if you close your laptop. When something needs testing on your machine, the coordinator starts a local agent. - Each Project has a set of files synced across machines, holding research, codebase knowledge, and your preferences. - It can subscribe to Slack channels, run on a schedule, or follow your PRs. | Comparison | Remote Control | Projects | |---|---|---| | Where it runs | Your computer | Cloud computer, calling a local agent when needed | | Role of the phone | View, reply | See the official docs | | Computer must stay on | Yes | Not for the cloud part | | Status | Shipped October 6 | Beta | ## What's new for cloud agents [The October 16 update](https://cursor.com/changelog/cloud-agent-canvases-and-environments) has three points. Cloud agents can generate interactive canvases and share them with your team by link. The Environments API is in beta, so you can create and manage cloud environments programmatically. A single environment now supports up to 300 repositories, and large environments load 5x faster. For the actual endpoints and parameters, see the official docs. ## Local agents need the network: how to debug a dropout Both ends of Remote Control depend on the computer being online. If the agent looks unresponsive from your phone, work through this in order: 1. Check [/en/down/cursor](/en/down/cursor) and [/en/status](/en/status) to see whether the problem is on the service side. For October's degradations, see [Cursor down or hitting 429s? October's degradations on the official status page](/en/blog/cursor-429-degraded-october-2026). 2. On the computer, open the [cdnping home page](/en) and run a test of first-connection and round-trip time for each API, to tell a local network problem from a route problem. For what the latency metrics mean, see [AI API latency breakdown](/en/blog/api-latency-connect-rtt-ttft). 3. Check whether the computer is asleep or unplugged, and whether "Keep this computer awake" is actually in effect. 4. If it still misbehaves, run through the [developer network self-check list (2026 edition)](/en/blog/developer-network-self-check-2026) to review your egress IP and DNS. Also, an agent running locally can execute commands on your machine. If you want an extra layer of isolation, see the approach in [Copilot local sandboxing is GA](/en/blog/github-copilot-local-sandboxing-ga). For Cursor's own permission settings, follow the official docs. ## Which one fits which scenario - A local task runs for a long time and you only need to approve something or add a line now and then: use Remote Control. - A big task needs to be split across several agents and run unattended for a long time: look at Projects, but remember it is still beta. - You want to manage cloud environments in bulk through code, with many repositories: look at the Environments API. ## Sources - [Cursor Changelog: Remote Control for local agents (2026-10-06)](https://cursor.com/changelog/remote-control-local-agents) - [Cursor Changelog: Projects (2026-09-10)](https://cursor.com/changelog/projects) - [Cursor Changelog: Cloud agent canvases and Environments API (2026-10-16)](https://cursor.com/changelog/cloud-agent-canvases-and-environments) - [Cursor Changelog](https://cursor.com/changelog)