Your AI Assistant Just Got a Sommelier

Your AI assistant knows a lot about wine. But it guesses.

Ask it "What wine goes with grilled salmon?" and you get a paragraph. But ask again tomorrow and the answer changes. There is no score. No explanation of why one wine works better than another.

That is because language models generate text. They do not calculate pairings. We built the calculation.

Introducing the SommelierX MCP Server

The SommelierX MCP server connects your AI assistant to a sommelier-grade wine pairing algorithm. One command:

npx @sommelierx/mcp-server

No signup. No API key for the free tier. 50 calls per day.

What It Does

7 tools available:

• Pair wine with a meal -- "What wine with ossobuco?"
• Pair wine with ingredients -- "I have salmon, dill, lemon"
• Reverse pairing -- "I have a Barolo. What should I cook?"
• Recipe URL pairing -- Paste a recipe link
• Group pairing -- Three courses, one wine
• Search -- 609 calculated wine styles

How It Works

Professional sommeliers helped profile 609 wine styles across 19 wine dimensions. The algorithm compares those profiles with 17 dish dimensions and returns a compatibility score from 0 to 100. The calculation is structured: the score comes out of fixed dimensions and weights rather than text written on the fly, so every dimension behind a score can be inspected.

Setup in 30 Seconds

Add this to your MCP client configuration:

{
  "mcpServers": {
    "sommelierx": {
      "command": "npx",
      "args": ["@sommelierx/mcp-server"]
    }
  }
}

Works with Claude Desktop, Cursor, Windsurf, and any MCP client.

Links

• npm: @sommelierx/mcp-server
• Docs: docs.sommelierx.com
• App: app.sommelierx.com