RDF-JS inference engine playground

Run generated-runtime inference in the browser with the bundled OWL 2 RL and SKOS Core rule profiles.

Use this in your own projects with the NPM library or find the code on Github.

Bundled example

Load one of the self-contained examples from the repository examples folder, then run it with OWL 2 RL and SKOS Core support enabled by default.

Advanced rule profile selection

All bundled N3 rule files are enabled by default. Turn individual profiles off when you want to isolate one ruleset, compare behavior, or debug which profile derives a result.

Background knowledge

Provide the ontology, alignments, context information... in TriG, Turtle, N-triples or N-quads.

Fetched when you run inference. The server must allow browser CORS requests.

Input data

Can be an RDF response (TriG, Turtle, N-triples or N-quads) or an RDF Messages log.

Use a named Eyeling persistent fact store per playground dataset, then reason over that store plus each next message. Checking this box means you tell the engine that each message is universally true, which is not the default behaviour. Check this box with care.
URL input is fetched when you run inference. RDF Messages URLs are streamed and emitted incrementally.
Advanced SHACL in/out shape hints

Optional trusted optimization contracts. These SHACL shapes are not used for validation; they help the generated runtime and per-input planner keep only rules, predicates, indexes, and joins relevant to the expected input and desired output.

Describes RDF Messages or input batches passed to inference.
Describes the inferred RDF you care about retaining.

Output

Newly inferred triples.