Putting research into practice
This page showcases all project resources and platforms developed in my research group.
The Agentic SPARQL Infrastructure: SPARQL-MCP & Catalogue-MCP
The Agentic SPARQL Infrastructure represents a paradigm shift in how Artificial Intelligence interacts with the Semantic Web. With SPARQL-MCP and the Catalogue-MCP, we enable AI agents to autonomously navigate the decentralized Web of Data - implementing the initial agentic vision for the Semantic Web.1 By leveraging the Model Context Protocol (MCP), agents can discover endpoints, explore schemas, and formulate federated SPARQL queries on-the-fly. This bridge between Large Language Models (sub-symbolic AI) and Knowledge Graphs (symbolic AI) empowers a new generation of Broad AI systems that reason over verifiable, up-to-date facts. We envision a comprehensive infrastructure of intelligent web agents and invite collaboration on:
- Complex orchestration & choreography, federated query architectures & cost models
- Integration with the Web of Things, access policies & SOLID infrastructure
- Further agentic protocols (e.g., A2A), agent integration in SPARQL endpoints
- Diverse Linked Data types (dumps, fragments), other data models (relational, graph models, GraphQL)
