Why MCP

Historically, Mobitru APIs were accessed via REST that required manual scripting and configuration.

The MCP for Enterprise AI plan transforms that workflow into conversational, AI-assisted automation, where every Mobitru capability is available directly via natural language. Instead of writing code or scripts, engineers can simply ask their assistant to perform test actions, for example:

"Reserve an iPhone 15, upload and install the Biometrics app, launch it, and record the screen.”
“Open the Apple homepage in Firefox, then repeat the same test in Chrome and WebKit.”

The MCP server interprets the intent, maps it to Mobitru actions, and executes them instantly — enabling frictionless collaboration between human testers, AI copilots, and Mobitru’s real infrastructure.

Each MCP server runs locally on the user’s computer and acts as a secure bridge between your AI assistant and Mobitru’s testing infrastructure.

When started (for example, using npx mobitru-mcp-server mobile), it exposes a catalog of structured “tools” — each representing a Mobitru feature or API action. These tools can be invoked manually or automatically by your AI assistant in response to natural language prompts such as:

“Start an Android device and open the mobitru.com website.”
“Open the Apple homepage in Chrome, Firefox, and WebKit.”

Supported MCP Servers

Depending on your workflow, you can launch one or both of the following servers:

• Mobile MCP — for managing and testing on real iOS and Android devices.
• Cross-browser MCP (Playwright MCP) — for running browser-based tests in Chromium, Chrome, Firefox, or WebKit (Safari).