Tauri Plugin: Model Context Protocol (MCP)

A Tauri plugin and MCP server that enables AI agents (Claude Code, Cursor, etc.) to interact with and debug Tauri desktop applications through standardized interfaces.

Author: Jarad DeLorenzo (@delorenj) Repository: https://github.com/delorenj/tauri-plugin-mcp Homepage: https://delorenj.github.io

Features

The Tauri MCP Plugin provides a comprehensive set of tools that allow AI models and external applications to interact with Tauri applications:

Window Interaction

• Take Screenshot: Capture images of any Tauri window with configurable quality and size
• Window Management: Control window position, size, focus, minimize/maximize state
• DOM Access: Retrieve the HTML DOM content from webviews windows

User Input Simulation

• Mouse Movement: Simulate mouse clicks, movements, and scrolling
• Text Input: Programmatically input text into focused elements
• Execute JavaScript: Run arbitrary JavaScript code in the application context

Data & Storage

• Local Storage Management: Get, set, remove, and clear localStorage entries
• Ping: Simple connectivity testing to verify the plugin is responsive

Quick Start

Building the Plugin

bun install
bun run build && bun run build-plugin

MCP Server Setup

cd mcp-server-ts
bun install
bun run build

Basic Configuration

Add to your Claude Code or Cursor MCP settings:

{
  "mcpServers": {
    "tauri-mcp": {
      "command": "node",
      "args": ["/path/to/tauri-plugin-mcp/mcp-server-ts/build/index.js"]
    }
  }
}

Documentation

• Setup Guide - Detailed integration instructions
• MCP Server - Available tools and configuration
• Troubleshooting - Common issues and solutions

Architecture

The plugin uses a multi-layer communication stack:

AI Agent (Claude Code/Cursor)
         ↓ (MCP Protocol)
    MCP Server (Node.js)
         ↓ (IPC/TCP Socket)
   Socket Server (Rust)
         ↓ (Tauri APIs)
   Tauri Application

Supports both IPC (Unix sockets/named pipes) and TCP connections for maximum flexibility.