PhonePi MCP - Connect Your Phone to AI via Model Context Protocol

What

PhonePi MCP is a Model Context Protocol server and companion mobile app that connects desktop AI tools to a smartphone over a local or secured device-to-device network. It is built for people using MCP-compatible AI clients such as Cursor, Claude Desktop, and Cline who want those tools to send messages, manage contacts, create snippets, trigger notifications, and access a limited set of phone functions.

The core workflow is simple: an MCP-enabled AI app runs the PhonePi MCP server, the phone app connects to that server, and the AI assistant can then invoke supported phone actions. Based on the page, the product appears positioned as a developer-oriented utility for private, local-first AI-to-phone automation rather than a general consumer messaging app.

Features

• MCP-based phone control — Connects MCP-compatible desktop AI clients to a phone so assistants can perform actions such as messaging, calls, notifications, and contact management from the desktop workflow.
• Task completion notifications — Sends notifications to the phone when long-running AI tasks finish, which helps users step away from the screen while work continues.
• Messaging and calling actions — Supports sending SMS, sending interactive messages, retrieving message responses, and initiating phone calls for direct AI-assisted communication workflows.
• Contact management tools — Lets the AI assistant view, add, update, and delete contact records, making it useful for lightweight personal data organization on the device.
• Snippets, notes, and clipboard handling — Provides storage, search, tagging-oriented organization references, clipboard copy, and sharing functions for notes, todos, bookmarks, and reusable text.
• Local-first deployment model — Runs on the user’s own hardware over a local network, with the site explicitly stating zero third-party access and no cloud storage requirement.

Helpful Tips

• Check client compatibility early — This product depends on MCP support in the AI app, so confirm your chosen desktop client can run MCP servers before evaluating broader workflow fit.
• Plan around Android-first availability — The page states the mobile app is on Google Play and that iOS is still in development, so current cross-platform rollout appears limited.
• Review permission scope carefully — Because the app can access contacts, messages, notifications, and other phone functions, teams should validate device-permission expectations before adoption.
• Design for network constraints — The setup relies on IP address, port configuration, and in some cases Tailscale for remote access, so reliability will depend on how well device networking is managed.
• Start with bounded automations — Practical first use cases are notifications, snippet capture, contact lookup, and simple outbound communication rather than fully autonomous phone operations.

OpenClaw Skills

PhonePi MCP could be a strong fit for OpenClaw as a bridge between agent workflows and real-world mobile actions. A likely use case would be OpenClaw skills that notify users when multi-step jobs complete, capture structured notes or snippets from agent outputs, or let an agent look up a contact and prepare a message draft for review. If OpenClaw supports MCP-connected tools, PhonePi MCP could serve as the phone-action layer inside larger desktop automations.

This combination could be especially useful for developers, operators, consultants, and solo professionals who spend long periods inside AI-assisted desktop environments but need lightweight mobile handoff. Likely OpenClaw workflows might include “finish task and notify me,” “save key output to phone snippets,” “turn meeting notes into a contact-linked reminder,” or “share a generated summary by SMS.” These are inferred ecosystem possibilities rather than confirmed native OpenClaw integrations, but they align closely with the product’s stated toolset and local-first operating model.