InspirationInspiration
Finding something interesting to do in a city is still surprisingly fragmented. Events are spread across venue websites, cultural institutions, clubs and organizers. Stadtoskop started as an attempt to turn this fragmented information into a visual and playful way to discover local culture.
While building it, another question became increasingly important: How should AI agents discover local culture?
Today, an agent often has to search the web, interpret different websites and reconstruct information that already exists as structured data.
WebMCP offers a much cleaner approach.
What it does
Stadtoskop is a local culture discovery platform starting in Leipzig, Germany. It organizes events, venues and categories into structured data while keeping links to their original public sources.
For the WebMCP Challenge, Stadtoskop adds an agent-accessible discovery layer.
Instead of scraping or visually interpreting the website, an AI agent can use WebMCP tools to search Stadtoskop directly — for example to find live music, cultural events or places matching a user's interests.
The goal is simple:
People explore the city visually. Agents explore the same city through structured tools.
How we built it
Stadtoskop already has a React Native frontend, a Node.js backend and a PostgreSQL/PostGIS database containing structured event and venue information.
The WebMCP integration exposes selected discovery capabilities as clearly defined tools. These tools connect the agent-facing web experience with the same underlying cultural data used by Stadtoskop.
This means the WebMCP layer does not create a second database or duplicate the discovery logic. It provides another interface to the existing Stadtoskop ecosystem.
Challenges
Local event data is messy. Names change, events repeat, venues move, and information can differ between sources.
A major part of Stadtoskop has therefore been about making discovery data reliable enough to be useful while still preserving the connection to official public sources.
The WebMCP challenge adds another interesting requirement: deciding which parts of that data model should become simple, useful tools for an AI agent instead of exposing the complexity of the entire system.
What's next
The WebMCP prototype is only the beginning.
The long-term idea is for Stadtoskop to become a local cultural discovery layer that can be used both by humans and AI assistants.
An agent could eventually answer questions such as:
“What can I do with my kids in Leipzig this weekend?”
or:
“Which bands playing in Leipzig during the next two weeks should I listen to?”
WebMCP provides a promising way to make that interaction direct, structured and useful.
Log in or sign up for Devpost to join the conversation.