Overview

Gaggimate MCP Server is an open-source project that enables AI agents to directly interact with espresso machines, specifically Gaggimate-modded Gaggia Classic machines. It transforms a general-purpose AI into a “barista coach” that understands extraction theory, tasting vocabulary, and brewing profiles.

GitHub Repository: julianleopold/gaggimate-mcp

A shot profile showing pressure decline, temperature curves, and shot notes.

A shot profile showing pressure decline, temperature curves, and shot notes.

What It Does

The MCP server provides AI agents with five core tools to analyze and optimize espresso brewing:

  • Read shot data: Access temperature, pressure, and flow rate metrics
  • Manage brewing profiles: Create and modify extraction profiles
  • Track feedback: Record tasting notes and shot ratings
  • Browse history: Analyze past shots for patterns
  • Diagnose issues: Troubleshoot connection problems
Claude calling the manage_profile tool to create a brewing profile on the Gaggimate.

Claude calling the manage_profile tool to create a brewing profile on the Gaggimate.

The Workflow

  1. Pull an espresso shot using your Gaggimate-modded machine
  2. AI prompts for your tasting feedback
  3. AI analyzes shot data combined with your notes
  4. AI suggests specific profile adjustments
  5. Repeat until you achieve the perfect extraction
Claude researching a new bag of beans — identifying origin, processing method, and extraction implications.

Claude researching a new bag of beans — identifying origin, processing method, and extraction implications.

Safety First

The system includes important guardrails:

  • Cannot remotely start or stop shots
  • Temperature limited to 25-100°C
  • Pressure capped at 0-12 bar
  • All AI-created profiles tagged with “[AI]” prefix

Technical Details

Built with:

  • Python 3.11+
  • Model Context Protocol (MCP): Anthropic’s standard for AI-tool integration
  • Gaggimate API: For espresso machine control and telemetry
  • Compatible with MCP clients like Claude Desktop

Why This Project?

Espresso brewing involves dozens of variables—grind size, temperature, pressure, flow rate—that interact in complex ways. This project explores whether AI agents can:

  • Understand espresso extraction theory
  • Translate tasting notes into actionable adjustments
  • Learn iteratively from shot data
  • Provide personalized brewing guidance

It’s an experiment in using AI for fine-grained hardware control and iterative optimization in a domain that traditionally requires years of experience.

Future Ideas

  • Multi-machine profile sharing
  • Automated grind size recommendations
  • Shot classification and pattern recognition
  • Integration with coffee bean databases
  • Community profile library

This project is open source and contributions are welcome. Check out the GitHub repository for installation instructions and documentation.