Anthropic Launches Claude Dispatch: Cross-Device Task Assignment with Persistent Conversations
Key Takeaways
- ▸Claude Dispatch enables seamless task assignment from mobile to desktop with persistent, context-aware conversations across devices
- ▸The feature leverages existing Cowork configurations, giving Claude access to local files, connectors, and plugins without requiring separate setup
- ▸Dispatch executes tasks asynchronously in the background, delivering results directly to users' devices when work is complete
Summary
Anthropic has introduced Claude Dispatch, a new research preview feature that enables users to assign tasks to Claude from their mobile devices while Claude executes them on their desktop computer. The capability maintains a single persistent conversation thread across phone and desktop, allowing Claude to retain context from previous interactions and work asynchronously on assigned tasks using local files, connectors, and plugins already configured in Cowork.
The feature is now available to Cowork Pro and Max plan subscribers and requires both the Claude Desktop app and Claude mobile app to be installed and up-to-date. Users can assign Claude tasks such as compiling data from local spreadsheets, searching email and Slack messages, creating formatted presentations, or organizing files—with Claude handling the work in the background and notifying users when tasks are complete.
Anthropically has emphasized important security considerations, noting that this capability creates a chain where mobile instructions can trigger real actions on a desktop computer, including file access, service interactions, and browser control. Users are advised to carefully review permissions granted to Claude in Cowork before enabling Dispatch, given the potential for mistakes or malicious instructions to have cascading, difficult-to-undo consequences.
- Security requires user vigilance: the mobile-to-desktop chain can trigger real file operations, service interactions, and browser actions with significant consequences
Editorial Opinion
Claude Dispatch represents a meaningful evolution in how users can interact with AI assistants, collapsing the friction between ideation on mobile and execution on desktop. The persistent conversation thread is particularly valuable, as maintaining context across devices reduces the cognitive load of re-briefing the model. However, Anthropic's transparent security warnings deserve serious user attention—the power to delegate tasks remotely cuts both ways, and the potential for cascading failures or prompt injection attacks is real. This is a feature that will likely appeal to power users comfortable managing permissions, but it underscores why AI safety practices at the user level matter just as much as algorithmic guardrails.
