I Built an MCP Server for KubeCon EU 2026: Here's How AI Agents Can Plan Your Conference

#ai #mcp #cloudnative #kubecon

KubeCon + CloudNativeCon Europe 2026 has 500+ sessions across 4 days, dozens of evening parties, 16 co-located events, and a venue the size of a small airport. Planning your week is a project in itself.

So I built an MCP server that lets AI assistants do it for you.

What's MCP?

Model Context Protocol is an open standard that lets AI models call external tools. Think of it as giving Claude (or any compatible agent) the ability to search databases, call APIs, and interact with real-world data — instead of just working from its training data.

What kubecon-eu-mcp Does

It exposes 12 tools, 4 resources, and 5 prompts that connect your AI assistant to live KubeCon data:

Session tools:

search_sessions — find talks by keyword, speaker, topic, or track
get_schedule — full schedule for any day
find_speaker — look up what someone is presenting
score_sessions — get personalized session rankings based on your role and interests (scoring rubric from kubecon-event-scorer)
detect_conflicts — check if your top picks overlap

Party tools:

find_parties — discover evening events scraped from conferenceparties.com
plan_party_route — optimized party-hopping route through Amsterdam

Logistics:

get_venue_info, get_hotel_info, get_travel_info — venue rooms, hotel blocks, transit from Schiphol, airline discount codes

Data comes live from the official sched.com iCal feed and conferenceparties.com, with in-memory TTL caching.

Try It

uvx kubecon-eu-mcp

Add to Claude Desktop (claude_desktop_config.json):

{
  "mcpServers": {
    "kubecon-eu-mcp": {
      "command": "uvx",
      "args": ["kubecon-eu-mcp"]
    }
  }
}

Then just ask naturally:

"What AI talks are on Wednesday?"
"Plan my evening on Tuesday — I want food and networking"
"Build me a 4-day itinerary as an SRE interested in eBPF and observability"
"Are there conflicts between the sessions I picked?"

The Meta Angle

KubeCon EU has a co-located event called Agentics Day: MCP + Agents on Monday — focused entirely on the Model Context Protocol and AI agents in cloud-native.

So yes, this is an MCP server to help you plan your attendance at the MCP event. We've achieved recursion.

How It's Built

The stack is intentionally simple:

Python with FastMCP (the official MCP Python SDK)
httpx for async upstream fetching
icalendar for parsing sched.com's iCal feeds
BeautifulSoup for scraping party listings
Pydantic for data validation

No database. No config files. Everything is fetched live and cached in memory. Install it, run it, done.

DEV Community