SPARQL Explorer

Connect to any SPARQL endpoint, discover the schema, explore classes and properties visually, and write queries with auto-complete.

-- Presets -- NFDI MatWerk DBpedia (large, lightweight mode) Wikidata (large, lightweight mode) Connect 50 items 100 items 300 items 1000 items All (slow)

+ − Fit DagreForceCircle ConcentricTree

PNG SVG

Classes

Properties

Query Builder

Template -- Select -- Count instances of a class Find by label (text search) All properties of a class Explore property values Describe a resource Sample triples Class -- Property -- Search text / IRI

Select a template and fill in the fields, or write your own SPARQL. Press Ctrl+Space in the editor for autocomplete (classes, properties, keywords with fuzzy search).

SELECT ?s ?p ?o WHERE { ?s ?p ?o } LIMIT 20

Run Query Visualize Results

Results

+ − Fit DagreForceCircle

PNG SVG

How to use

1. Select a preset or enter your SPARQL endpoint URL
2. Click Connect — the tool discovers classes, properties, and hierarchy
3. Explore the schema graph (abstract model showing classes and relationships)
4. Click classes/properties in the sidebar to select them for queries
5. Choose a query template or press Ctrl+Space for autocomplete
6. View results as a table or visualize them as a graph

Share via URL

Link directly to an endpoint:

https://ise-fizkarlsruhe.github.io/ontoink/sparql-explorer/?endpoint=https://nfdi.fiz-karlsruhe.de/matwerk/sparql

How it handles large endpoints

For knowledge graphs with millions of triples (DBpedia, Wikidata), ontoink uses adaptive discovery:

1. Fast probe — checks endpoint responds (1 triple, 10s timeout)
2. Class discovery with counts — tries GROUP BY with 15s timeout
3. Fallback to lightweight mode — if counts time out, uses DISTINCT (no counting, much faster)
4. Property discovery — tries domain/range detection, falls back to distinct properties
5. Label fetching — batch lookup for the top 80 IRIs

This means any endpoint works — small ones get full statistics, large ones get schema structure without counts.

Requirements

• The SPARQL endpoint must support CORS (cross-origin requests)
• Most public endpoints (DBpedia, Wikidata, NFDI) support this
• For private endpoints, configure CORS headers on your server