2026-11-03 –, Beavis
AI assistants can now help administer GeoServer and MapStore via Model Context Protocol servers, publishing layers, tuning caches, diagnosing issues, with guardrails to keep them safe. An early look at what we're building and learning.
Administering GeoServer and MapStore still means clicking through admin UIs or scripting against REST APIs — even for routine tasks like publishing a layer, tuning a cache, or diagnosing slowness. What if you could just ask?
We're building Model Context Protocol (MCP) servers for GeoServer and MapStore, designed to work with today's major AI assistants, letting them help with real administrative and configuration tasks through natural language, while setting up guardrails preventing the agent from going ballistic on your deploy.
This talk shares an early look at the project: what we're learning about designing AI-friendly tools for GIS server administration, and the open questions around security, efficiency, and trust when giving an AI a hand on production geospatial infrastructure. Come see where we are, what's working, and where we'd love community input.
Open source enthusiast with strong experience in Java development and GIS. Personal interest range from high performance software, managing large data volumes, software testing and quality, spatial data analysis algorithms, map rendering.
Full time open source developer on GeoServer and GeoTools, regular presenter at F0SS4G.
Received the Sol Katz's OSGeo award in 2017.
Simone Giannecchini is the founder and Technical Director of GeoSolutions, a leading contributor to open-source geospatial technologies such as GeoServer, MapStore, and GeoNode. With over two decades of experience in large-scale, standards-based system design, he has led deployments of spatial data infrastructures and 3D applications for public and private sector clients globally.