Purpose: mobile-mcp is a Model Context Protocol (MCP) server for mobile automation, development, and data scraping. It provides a platform-independent interface, enabling agents and LLMs to operate native apps and devices without separately mastering iOS or Android knowledge, and is suitable for scenarios such as testing, data entry, and multi-step user flow automation.
Main capabilities: supports iOS and Android simulators, emulators, and real devices; interacts through structured accessibility snapshots or screenshot-based coordinate taps; provides device management (list devices, screen size, orientation, foldable devices, GPS location, clipboard), app management (list/launch/terminate/install/uninstall apps), screen interaction (screenshots, list UI elements, tap, double tap, long press, swipe, screen recording), input and navigation (enter text, key presses, open URL), logs and crash reports, and batch commands. It also supports using cloud real devices through Mobile Next Cloud.
Usage: requires installation of Xcode command line tools and Android Platform Tools, and works with MCP-compatible clients (such as Claude Code, Codex, Gemini, GitHub Copilot, etc.). Most tools can use the standard configuration, with the command npx -y @mobilenext/mobile-mcp@latest. By default it runs over stdio, and it can also use --listen to start a Streamable HTTP server bound to localhost:3000, with clients connecting to http://<host>:3000/mcp; MOBILEMCP_AUTH can be set to require a Bearer token. After configuration, have the agent list devices to verify.
Limitations: the mobile platform SDK must be configured in advance; the iOS simulator must run on macOS/Linux, the Android emulator supports Linux/Windows/macOS; real devices require the corresponding platform tools and drivers. Anonymous usage telemetry is collected by default and can be disabled with MOBILEMCP_DISABLE_TELEMETRY. Specific pricing, costs, and performance data are not yet provided.
