Cursor Remote Control

You can check a Cursor agent and send its next instruction from your iPhone while the work continues on your laptop. For a developer already using local agents, that makes a short reply possible without returning to the desk. Cursor's October 6, 2026 instructions cover pairing the devices; your computer still needs to stay on and online. Cursor's release note.

Set Up Cursor Remote Control on Your iPhone

Pair the phone while you are still beside the computer. Desktop approval is part of the setup, so do this before leaving an agent unattended.

1. Install the Cursor iOS app and sign in. Use the download link on Cursor's mobile page. Sign in to the account you use on your computer.
2. Select your computer. Computers associated with your account appear automatically in the app. Tap the one running your local agent.
3. Approve pairing on the computer. Accept the pairing request in the Cursor desktop app.
4. Open a local agent on your iPhone. After pairing, choose an agent from the list to see its work or send a message.

That is the procedure in the October 6 announcement. For a first check, try a bounded message such as: “Summarize the current task and tell me what decision you need next.” Confirm that the reply is about the task you expected before sending further instructions.

Release scope matters. Cursor's mobile reference describes a cloud-agent handoff and different admin controls. It conflicts with the October 6 announcement, which explicitly says cloud agents are unnecessary. This guide follows the dated October 6 pairing instructions. The two procedures should not be combined into one setup checklist.

Keep the Laptop Reachable

Enable Keep this computer awake under Remote Control in Cursor's desktop settings. Cursor specifies that the computer must be plugged in with the lid open. It must also remain on and online. Awake requirements.

Treat that as your departure check: power connected, lid open, network working. Closing the laptop and putting it in your bag does not meet Cursor's stated conditions.

A useful first trial is to step into another room and ask the existing agent for a status summary. Leave a noncritical task running while you check that the setup behaves as expected.

Availability and Enterprise Setup

Remote Control is enabled by default outside Enterprise organizations. Enterprise admins enable it at Org settings > Security and identity > Remote control. The October 6 release does not require cloud agents and names no plan restriction for this Remote Control flow. Cursor availability and admin instructions.

The native app identified by Cursor is for iOS. Android support is not established by the announcement or mobile product page.

What Runs on the Phone, Laptop or Cloud?

Your phone is the place where you give directions. The computer remains the place where the local agent works. Think of a phone call to someone still at your workbench: changing where you speak does not move the workbench.

Cursor's mobile page presents both Remote Control and Cloud Agents. That shared app can make the distinction easy to miss. Cloud Agents use separate virtual machines, which are computers hosted in the cloud with their own development environments. Cursor says those agents do not require your local computer to stay connected to the internet. Cursor Cloud Agents.

Claude Code connects a session. Anthropic documents claude remote-control for starting the connection from a terminal and /remote-control within an existing session. A session URL or QR code takes you to the conversation, and the Claude app or browser can continue it. The computer and Claude process must stay running. That is a different entry point from selecting a computer and approving Cursor's pairing request. Claude Code Remote Control.

Codex connects a host through the desktop app. OpenAI's current guide starts in the ChatGPT desktop app at Settings > Connections > Control this Mac or PC, followed by Set up or Add and QR pairing. Desktop and mobile use the same account and workspace. The host needs to remain awake and online; mobile setup does not start from the Codex CLI or IDE extension. OpenAI also documents remote approvals and reviewing diffs, the changes made to files. OpenAI remote connections.

If you are choosing a tool rather than configuring an existing Cursor workflow, use the mobile coding-agent comparison. For the broader editor decision, see the Cursor review.

What Does Cursor Remote Control Cost?

Cursor's October 6 announcement names no Remote Control plan restriction, and its pricing page lists no separate Remote Control charge. That does not promise unlimited agent usage. The distinction is access to the connection versus the model usage consumed by the agent.

The default monthly view on cursor.com/pricing, checked October 9, 2026, shows these headline options:

These are subscription prices, not a list of required Remote Control upgrades. Cursor lists limited agent requests for Hobby and explains that on-demand model usage can be billed after included usage is consumed. Prices exclude applicable taxes. Pro+, Ultra and Teams Premium selectors also appear on the page; the table reproduces its default selections. Official pricing. The site's Cursor pricing guide covers the wider purchasing decision.

The useful business calculation is time between decisions. As an illustrative scenario, suppose your own work log shows four weekly returns to the laptop taking 15 minutes each. Replacing each with a two-minute phone exchange would reduce that time from 60 minutes to eight, a potential 52 minutes per week. These are assumptions to test, not measured Cursor results. Additional agent usage and the cost of keeping the host running still belong in your calculation.

Six Useful Ways to Steer an Existing Agent

The strongest use is a short decision that keeps a well-scoped task moving. The examples below are proposed workflows, ranked by likely value to a developer who already has the agent running locally.

These examples use ordinary messages. A prompt asking an agent to preserve scope is an instruction, not an enforcement mechanism. Review the resulting code and test evidence before treating the task as finished.

Two Things Worth Building Around This Workflow

The strongest opportunity is a team handoff kit. Small agencies could pay for a repeatable way to start local tasks, ask for concise progress reports and resume review at a desk. DataForSEO's US keyword estimate, checked October 9, is 260 monthly searches for “cursor remote control.” That is evidence of interest in the workflow, not a count of buyers.

The smallest useful product could be a repository template with task briefs, a standard “changed files, tests, next decision” report, mobile-friendly prompt snippets and a pairing checklist. Start with a team whose existing tasks regularly stall on one person's reply. The catch is straightforward: Cursor already supplies the connection, so a generic remote-control wrapper adds little. The value would have to come from improving that team's handoffs. This could be delivered without assuming an undocumented Remote Control API exists.

A laptop status helper is a narrower second bet. Developers leaving an agent at the desk could use a small local utility that reports power and network status, explains the open-lid condition, and reminds them to check Cursor's awake setting. DataForSEO estimates 210 monthly US searches for “coding from phone.” This is an adjacent interest signal; it does not prove demand for a paid readiness tool.

An initial version could be a local status panel and a pre-departure checklist using documented operating-system interfaces. It should not claim to verify Cursor pairing without a supported integration. The catch is platform maintenance and limited differentiation: Cursor already offers the central awake setting. Validate repeated failures in a real team's routine before expanding the product.

Both opportunities are deliberately small. Search volume does not justify building another general-purpose coding platform.

If Your Computer or Agent Does Not Appear

Start with the prerequisites that match the documented flow. This is a diagnostic checklist inferred from those requirements, not an official error-code guide.

• No computer listed: check that desktop and phone use the intended account and that the host is available.
• Pairing has not completed: return to Cursor on the computer and check for the pairing request.
• An Enterprise account cannot connect: have the admin check the organization setting above.
• The connection stops after you leave: check power, lid position, sleep and network availability before assuming an agent problem.
• No expected local agent: confirm the task is running on the computer you selected.

Remote access does not make a sleeping laptop a dependable unattended worker. Choose a cloud execution workflow when laptop independence is the requirement. Local execution also does not, by itself, establish an offline or device-only data-processing guarantee.

Questions About Coding From Your Phone

Can I do coding using my phone?

For this workflow, you can check and direct a Cursor agent from the iOS app while the agent works on your computer. It is useful for short instructions and progress checks; the local development environment remains on the host. Cursor mobile.

Which app is best for coding on a phone?

If your active tasks already live in Cursor on your laptop, start with Cursor's own iOS app. If they live in Claude Code or Codex, use the corresponding remote connection. Choose according to the workflow you need to continue, rather than adding a second agent solely for phone access.

Is Cursor Remote Control available on Android?

Cursor's October 6 announcement and mobile product page identify iOS. They do not establish Android support for this setup. Cursor mobile.

Where do I download the Cursor app for iPhone?

Follow the iOS download link on Cursor's official mobile page, then sign in to the account associated with your computer.

Your Monday Move

Before leaving your desk, pair the iPhone, configure the awake setting and open a small local task you can check safely. Use one phone message to resolve its next decision, then review the work on your laptop. Measure whether that exchange saved a return trip or simply added another interruption. That result tells you whether to make remote check-ins part of your normal routine.

If your team needs a repeatable agent handoff workflow, I build AI agent systems around the way your team works.