DEV Community

Cover image for MCP Server to generate custom QR codes directly in Cursor and Claude
Evgenii Zinner
Evgenii Zinner

Posted on

MCP Server to generate custom QR codes directly in Cursor and Claude

#ai #mcp #qr #agents

If you use Cursor, Claude Desktop, or Cline as your daily driver, you know how annoying it is to leave your editor just to generate simple project assets.

Last week, I was testing print layouts for a physical marketing project that needed scannable codes. I was tired of opening browser tabs, fighting spammy online generators that lock vector downloads behind paywalls, or getting forced into a subscription just to get an SVG.

So, I built an MCP (Model Context Protocol) server for my QR Generator.

Now, my AI assistant can generate high-resolution, customized QR codes (including clean vectors, custom shapes, and embedded logos) directly inside my IDE.

Here is how it works, what it does, and how to hook it up to your environment.

What it does (and why standard PNGs suck)

Most basic QR generators output low-resolution PNGs. If you try to print them on packaging or signage, they pixelate and scan terribly under bad lighting.

To solve this, the MCP server exposes a single, highly flexible tool: generate_qr_code. It lets you control:

  • Format: Choose between standard png, svg (perfect for print), or raw json.
  • Design Physics: Custom dot styles like circle, hex (hexagon), and diamond instead of boring old squares.
  • Color & Branding: Direct primary hex color mapping (e.g., #00f0ff) and custom logo uploads inside the data grid.

Setting it up in your IDE

Because the server runs over Server-Sent Events (SSE) hosted on our cloud backend, you don't need to install local runtimes or configure Python/Node environments.

To add it to your Claude Desktop or Cursor configuration, open your claude_desktop_config.json (or your Cursor MCP settings) and paste this snippet:

{
  "mcpServers": {
    "ez-qr-generator": {
      "url": "https://ez-qr-generator.com/mcp",
      "transport": "sse"
    }
  }
}
Enter fullscreen mode Exit fullscreen mode

Once saved, your AI agent will automatically detect the connection.

How to use it in practice

You don't need to write any JSON arguments manually. Just tell your assistant what you want in plain English. For example, if you are working in Cursor, you can prompt your agent:

"Hey, generate a hexagonal green QR code pointing to https://ez-qr-generator.com and place it in my assets folder as a SVG."

The assistant calls the server, processes the base64 or SVG output, and writes the file directly to your workspace. No context switching, no signups, and no paywalls.

If you want to test the raw JSON-RPC endpoint yourself or see the official registry, you can check out the Smithery page.

Give it a spin, and let me know if you run into any schema or connection issues!

Top comments (0)