Claude Dispatch: Phone-to-Desktop Remote for Cowork
Anthropic launches Claude Dispatch for phone-to-desktop remote control. Pair mobile with Mac via QR code, send tasks from iPhone to desktop.
Pairing Method
Average Setup Time
Task Execution
Co-Work Pattern
Key Takeaways
The most persistent limitation of mobile AI is the gap between what you can ask from a phone and what Claude can actually execute in a full desktop environment. Conversation, queries, and short content generation work well on mobile. But file operations, code editing, multi-tab research, and complex multi-step workflows require a desktop — and requiring you to be at your desk defeats the purpose of an AI that should follow your workflow, not dictate it. Claude Dispatch is Anthropic's answer to this problem.
By pairing your iPhone with a Mac through a QR code, Dispatch turns your phone into a command interface for your desktop AI. You describe a task from wherever you are; Claude executes it on the Mac using full desktop capabilities and notifies you when it is done. This co-work pattern — submit from mobile, execute on desktop, receive asynchronously — matches how knowledge workers actually move through hybrid and remote work environments. For teams already exploring how AI and digital transformation can reshape productivity workflows, Dispatch represents a meaningful architectural shift in how AI assistance is delivered.
What Is Claude Dispatch
Claude Dispatch is a feature within Anthropic's Claude platform that creates a direct pairing between a mobile device and a desktop running Claude. Once paired, the mobile device can submit tasks that Claude executes in the desktop environment, rather than on mobile infrastructure. The desktop becomes the AI's operating environment; the phone becomes its task input channel.
The distinction from standard remote access tools like Screen Sharing or remote desktop applications is important. Dispatch is not about remotely viewing or controlling your desktop screen from your phone. It is about delegating tasks to Claude on your desktop — you describe what you want in natural language, Claude figures out how to accomplish it using the desktop environment, executes the work, and returns results. You interact with the outcome, not with a live desktop session.
Your iPhone is the task submission interface. Describe what you need in natural language, attach context if needed, and send. No screen sharing, no remote viewing — just task delegation in plain language.
Claude on your Mac receives the task and executes it using full desktop capabilities: local file access, terminal, browser, applications, and the compute resources of your Mac rather than mobile infrastructure.
Task completion notifications arrive on your phone when Claude finishes. Results, created files, and execution summaries are available in the Claude interface on either device. No need to wait at the desktop.
The co-work framing in Anthropic's product messaging reflects a deliberate philosophy: AI should work alongside you on tasks while you continue other activities, not require your constant attention. This connects to Anthropic's broader direction in agentic AI products. For context on related capabilities, see our coverage of Anthropic's co-work recurring task automation for the recurring workflow dimension of the same product direction.
QR Code Pairing and Setup Guide
Setup is intentionally minimal. Anthropic designed the pairing process to take under a minute, and the connection persists across sessions without requiring the QR scan to be repeated. The steps below assume you have Claude installed on both your Mac and iPhone and are signed into the same Anthropic account on both devices.
Open Claude on Mac
Navigate to Settings → Dispatch → Pair Device. Claude generates a unique QR code tied to your account and this Mac.
Scan with Claude iOS
Open the Claude app on your iPhone, tap the Dispatch icon in the top navigation, and select Pair New Desktop. Point your camera at the QR code on your Mac screen.
Confirm the Pairing
Both devices display a confirmation prompt showing the device names. Confirm on both devices to establish the authenticated connection.
Set Permissions
macOS will prompt for permissions for the capabilities Claude Dispatch will use: file system access (specific folders you choose), terminal access, and browser access. Grant only what you need.
Test With a Simple Task
Send a simple task from your phone such as 'List the files in my Downloads folder and tell me which ones are largest.' Confirm it executes correctly before using Dispatch for important work.
Network note: For the initial QR pairing, both devices should be on the same Wi-Fi network. After pairing, Dispatch can relay tasks over the internet when your phone and Mac are on different networks, though local network communication is faster and preferred when available.
You can pair multiple Macs to a single iPhone — for example, your work MacBook Pro and a home Mac Studio — and select which desktop to dispatch tasks to when submitting from mobile. Each paired desktop appears as a named device in the Dispatch interface on your phone. Managing devices, revoking access, and reviewing Dispatch history are all available in your Anthropic account settings.
Task Types Supported on Desktop
Claude Dispatch routes tasks to the desktop based on their requirements. Tasks that benefit from desktop capabilities are dispatched to the Mac; tasks that are conversational or self-contained are handled directly in the mobile app. The classification happens automatically, so you do not need to explicitly specify where a task should run — though you can override the routing if needed.
Read, write, reorganize, and analyze local files. Compress folders, rename batches of files, extract data from documents, search across file contents, and generate summaries of local document collections.
Edit source files, run scripts, execute tests, debug errors, and commit changes to Git repositories on your local machine. Claude operates on your actual codebase, not a copy or sandbox environment.
Multi-tab browser research, synthesizing information from multiple sources, downloading and processing documents, and compiling research into structured outputs saved locally on your Mac.
Interact with desktop applications through AppleScript and Accessibility APIs. Create calendar events, update spreadsheets, compose emails in desktop clients, and interact with applications that have no API.
The task routing intelligence is built on Claude's understanding of what each task requires. A task like "summarize this document" can run on mobile if you paste the text, but "summarize the Q1 budget spreadsheet in my Documents folder" requires desktop file access and routes to your Mac automatically. The mobile app shows a visual indicator when a task has been dispatched to desktop so you know Claude is working on it there.
For teams using Claude extensively for enterprise document work, the desktop capabilities matter significantly. See our guide on Claude with Excel and PowerPoint for enterprise workflows for how desktop file access enables structured document automation that mobile-only AI cannot match.
Privacy Model and Security Architecture
For professional users, the privacy implications of giving an AI system access to local files and desktop applications are significant. Mistaking the convenience of Dispatch for a loosening of data privacy would be a mistake. The architecture is designed with privacy as a primary constraint.
Task execution happens on your Mac, not on Anthropic's cloud. File contents processed during a task are not uploaded to external servers beyond what standard API calls involve. The Mac is the execution environment, not a relay to cloud processing.
macOS permissions govern what Claude Dispatch can access. You choose which folders are accessible, whether terminal access is granted, and which applications Claude can interact with. Permissions can be changed at any time in System Settings.
Recommended permission scope: Grant Claude Dispatch access only to the specific folders relevant to your work — project directories, documents folder — rather than full disk access. This follows the principle of least privilege and limits the blast radius if a task goes unexpectedly wrong.
Enterprise consideration: Organizations subject to compliance requirements should review Anthropic's data processing addendum for Dispatch before deploying in environments with regulated data (PII, HIPAA-covered data, financial records). The local execution model reduces but does not eliminate data handling obligations.
Device authentication uses your Anthropic account credentials plus device-level cryptographic keys established during QR pairing. Losing your phone does not automatically grant the finder access to your desktop — they would need to authenticate with your Anthropic account credentials as well. Revoking a paired device immediately terminates its ability to dispatch tasks, even if the device is offline at the time of revocation.
Co-Work Workflows for Remote and Hybrid Teams
The co-work pattern that Dispatch enables maps well to how distributed teams actually operate. Work is not synchronous and desktop-bound — people move between meetings, commute, work from client sites, and need to initiate desktop-class work without returning to a fixed workstation. The following workflows illustrate where Dispatch changes the practical experience of remote and hybrid work.
During a five-minute break between back-to-back calls, send a task from your phone: "Process the new data export in my Downloads folder and update the client report template with the latest figures." Claude handles it while you take your next meeting.
Queue complex tasks during your commute that Claude executes on your desktop before you arrive. Research briefs, document drafts, code refactoring, and data processing tasks are ready when you sit down, reducing startup friction at the beginning of the workday.
Working at a client location without your laptop, you can dispatch tasks that require your desktop environment — code changes, document edits, report updates — without remote desktop latency or the security concerns of accessing work systems from a client machine.
From your phone before a morning meeting, dispatch tasks that aggregate data, generate summaries, and prepare presentation materials on your desktop overnight. Materials are ready without sitting at your desk before the meeting starts.
Dispatch vs Existing AI Workflows
Understanding where Claude Dispatch fits relative to existing tools and AI workflows helps set appropriate expectations and identify where it adds genuine value versus where existing approaches are sufficient.
Mobile Claude handles conversational tasks and simple queries directly. Dispatch adds desktop execution for tasks that need local file access or application interaction. The mobile app automatically routes to Dispatch when desktop capabilities are needed.
Remote desktop gives you full interactive control of your screen from a mobile interface, which is powerful but awkward on a small screen. Dispatch delegates work to Claude rather than giving you manual remote control — you describe outcomes, not click sequences.
Shortcuts and Automator handle well-defined, repeatable workflows you have pre-configured. Dispatch handles novel, one-off tasks described in natural language that you cannot anticipate and pre-program. Both have a place in a productive Mac workflow.
Custom API integrations provide programmatic automation for high-volume, predictable workflows. Dispatch handles ad hoc tasks that don't warrant building a custom integration. Engineers use both: automation for known workflows, Dispatch for novel ones.
Enterprise Use Cases and Productivity Gains
The productivity value of Dispatch comes from reducing context switching and wait times in knowledge work. Rather than mentally parking a task until you are at your desktop, you can delegate it immediately and continue with other work. Across a day of typical knowledge work, this eliminates significant context-switching friction.
Dispatch research and first-draft generation tasks while conducting client interviews. Files are ready to review when the interview concludes. Content pipeline throughput increases without additional desktop time.
Queue code review prep, test execution, and documentation updates from mobile during standups and reviews. Dispatch runs the tasks on the development machine while the engineer participates in meetings.
Account managers and consultants at client sites can dispatch deliverable preparation tasks back to their desktop without interrupting the client engagement. Follow-up materials are ready before they return to the office.
The measurable gains depend on task density and mobility patterns. Teams with high meeting loads and frequent off-desk time see the greatest benefit — knowledge workers who spend four or more hours per day away from their desk have more opportunities to use Dispatch effectively. For organizations building systematic AI workflows, Dispatch complements structured automation by handling the novel, one-off tasks that do not warrant building dedicated automation pipelines.
Current Limitations and Roadmap
Claude Dispatch is a new capability and ships with limitations that define where it fits best in the current release. Understanding these boundaries helps set realistic expectations and avoid workflow designs that will not work as intended.
iPhone and Mac only at launch: Android and Windows are on the roadmap. Cross-platform pairs (Android to Mac, iPhone to Windows) are not currently supported. Organizations with mixed device environments will need to plan for partial rollout.
Long-running task limitations: Tasks that require hours of continuous execution may be interrupted by Mac sleep, system updates, or network disruptions. For long-running automation, dedicated server-side orchestration is more reliable than Dispatch running on a personal Mac.
Application compatibility varies: Desktop application interaction depends on Accessibility API support and AppleScript compatibility. Some applications, particularly those with restrictive sandboxing, may not be accessible to Dispatch's automation capabilities.
No collaborative dispatching: Dispatch is a one-to-one feature between your phone and your Mac. Team workflows where multiple people dispatch to a shared machine are not supported. Each team member operates their own Dispatch pair.
Offline mobile queuing is limited: Submitting tasks when your phone lacks internet connectivity requires cached session data. Fully offline task submission is not supported in the initial release.
Anthropic's roadmap signals expansion of Dispatch to additional platforms and deeper integration with enterprise identity systems. The co-work direction points toward a model where AI agents work continuously in background desktop environments while humans remain in control of task initiation from any device. Teams that adopt the workflow patterns Dispatch enables now will be positioned to scale those patterns as the capabilities expand.
Conclusion
Claude Dispatch addresses a real gap in AI-assisted work: the mismatch between where people are and where the computing environment needed to do useful AI work lives. By making your Mac the execution environment and your phone the command interface, Dispatch enables co-work patterns where AI handles desktop-class tasks asynchronously while you stay mobile. The QR pairing setup keeps friction low, the permission model keeps privacy manageable, and the async notification flow keeps the pattern from demanding your attention.
The initial platform limitations — iPhone and Mac only, no collaborative dispatching, variable application compatibility — define where Dispatch fits best in the current release. For Apple ecosystem users with high meeting loads and significant off-desk time, it is immediately useful. For organizations evaluating Anthropic's direction in agentic and co-work AI, Dispatch is a clear signal of where the product is heading: AI that works alongside you continuously, not just when you are sitting in front of a screen.
Ready to Build Smarter AI Workflows?
Mobile-to-desktop AI co-work is one piece of a broader productivity transformation. Our team helps businesses design and implement AI workflows that deliver measurable efficiency gains across every part of the organization.
Related Articles
Continue exploring with these related guides