EHDS Linked Health Data Portal

120

Synthetic patients

3.37M

RDF triples

Clinical cohorts

MCP tools

Benchmark queries

MCP Connector

Connect any MCP-compatible AI agent — Claude, mcphost, or Python SDK — to query the data space.

Connect via SSE

SPARQL Explorer

Write and run SPARQL queries against the HealthDCAT-AP catalogue and FHIR-on-RDF clinical data.

Open SPARQL UI

RAG Vector Store

Download the pre-built ChromaDB vector store for RAG-condition evaluation.

Download ChromaDB

Datasets

graph/diabetes-cohort — 48 patients, Type 2 Diabetes (SNOMED 44054006), 1.35M triples
graph/hypertension-cohort — 50 patients, Essential Hypertension (SNOMED 59621000), 1.19M triples
graph/metabolic-syndrome-cohort — 22 patients, Metabolic Syndrome (SNOMED 237602007), 833K triples
graph/catalogue — HealthDCAT-AP Release 5 metadata catalogue, 144 triples

Quick Start

Claude.ai: Settings → Connectors → Add custom connector

https://mcp.linkeddata.es/sse

mcphost:

echo '{"mcpServers":{"ehds":{"url":"https://mcp.linkeddata.es/sse"}}}' > ~/.mcp.json
mcphost -m ollama:llama3.2

Python:

from mcp import ClientSession
from mcp.client.sse import sse_client

async with sse_client("https://mcp.linkeddata.es/sse") as (r, w):
    async with ClientSession(r, w) as session:
        await session.initialize()
        result = await session.call_tool("ehds_list_datasets", {})

Citation

Manab et al. (2026). EHDS Linked Health Data Portal.
ISWC 2026 Resource Track. https://mcp.linkeddata.es