Skip to main content
Glama
cafe-mcp-server

cafe-mcp-server

by hanluOMH

cafe-mcp-server

A minimal Python MCP server that recommends coffee from local static data.

This demo does not call external APIs and does not require an API key.

Tools

  • list_coffee_menu: returns the full static menu.

  • recommend_coffee: recommends one drink from mood, milk, caffeine, and temperature preferences.

  • explain_recommendation: explains a selected drink against the same preference inputs.

Related MCP server: mcp-coffee

Install

cd /Users/hanlufeng/Desktop/AI-INSIGHT-REPORT/Ceilia/cafe-mcp-server
python -m venv .venv
source .venv/bin/activate
python -m pip install -e ".[dev]"

Run

Local stdio MCP server:

cafe-mcp-server

Equivalent module form:

python -m cafe_mcp_server.server

Cloud Run / HTTP mode:

PORT=8080 python -m cafe_mcp_server.server

When PORT is set, the server automatically uses Streamable HTTP, binds to 0.0.0.0:$PORT, and exposes MCP at:

/mcp

You can also force the transport explicitly:

MCP_TRANSPORT=streamable-http PORT=8080 python -m cafe_mcp_server.server
MCP_TRANSPORT=stdio python -m cafe_mcp_server.server

For Cloud Run repository deployment with Buildpacks:

  • Build type: Buildpacks

  • Build context directory: /

  • Entrypoint: leave blank if Buildpacks uses the included Procfile, or set python -m cafe_mcp_server.server

  • Function target: leave blank

Example MCP config

{
  "mcpServers": {
    "cafe-recommendation": {
      "command": "python",
      "args": [
        "-m",
        "cafe_mcp_server.server"
      ],
      "cwd": "/Users/hanlufeng/Desktop/AI-INSIGHT-REPORT/Ceilia/cafe-mcp-server"
    }
  }
}

If the package is installed into the same Python environment used by the MCP client, the console script can be used instead:

{
  "mcpServers": {
    "cafe-recommendation": {
      "command": "cafe-mcp-server"
    }
  }
}

Development

pytest
python -m compileall src tests
PYTHONPATH=src python -c "from cafe_mcp_server.recommender import recommend_coffee; print(recommend_coffee(mood='smooth iced', prefer_milk=False)['name'])"

Command-line MCP tool test

After installing the project, call the MCP server through a stdio client:

python scripts/call_tool.py recommend_coffee --args '{"mood":"smooth iced","prefer_milk":false,"caffeine":"high","temperature":"cold"}'

Expected result includes:

{
  "id": "cold_brew",
  "name": "Cold Brew"
}

Available Tools

3 tools
explain_recommendationC

Explain a coffee choice against the same preference inputs.

Inputs: coffee_id (required), mood, caffeine, prefer_milk, +1 more. Returns: result

list_coffee_menuA

List all available coffee menu items.

Inputs: no named parameters. Returns: result

recommend_coffeeA

Recommend a coffee from mood and simple preferences.

Args: mood: Free-text preference such as "smooth iced" or "quick energy". prefer_milk: True for milk drinks, False for black coffee, None for either. caffeine: Optional caffeine level: low, medium, or high. temperature: Optional drink style: hot or cold.

Inputs: mood, caffeine, prefer_milk, temperature.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.5/5.0

    Scored across 3 tools

    Disambiguation4/5

    The tools are mostly distinct: listing the menu, recommending a coffee, and explaining a recommendation. There is slight potential for confusion between recommend_coffee and explain_recommendation, but their purposes are clear enough to avoid misselection.

    Naming Consistency5/5

    All tool names follow a consistent verb_noun snake_case pattern: list_coffee_menu, recommend_coffee, explain_recommendation. This makes the API predictable and easy to navigate.

    Tool Count4/5

    With 3 tools, the server is on the smaller side but well-scoped for its purpose. Each tool serves a distinct function in the coffee recommendation workflow, and the count does not feel excessive or incomplete.

    Completeness4/5

    The server covers the core domain of browsing, recommending, and explaining coffee choices. It lacks advanced features like detailed item descriptions or user feedback loops, but these are not essential for a basic cafe menu assistant.

    Maintenance

    ActivityInactive
    ResponsivenessOutdated data

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      A
      quality
      D
      maintenance
      MCP server for controlling Meticulous espresso machines via Claude and other AI clients.
      22
      105 npm
      MIT
    • A
      license
      A
      quality
      D
      maintenance
      An MCP project that orchestrates coffee orders by coordinating a Barista server and machine servers (grinder, brew unit, steamer, dispenser) through an LLM.
      7
      MIT
    • A
      license
      A
      quality
      C
      maintenance
      MCP server for querying Percolate's specialty coffee database with read-only access to 1,100+ coffees, tasting profiles, brew recipes, pairings, and recommendations.
      8
      55 npm
      MIT