Use SPARQL, Cypher, or GQL to ask questions of the same graph.
Legra exposes graph queries through SPARQL, Cypher, and GQL. The workspace remains the source of knowledge whichever supported language you use.
Choose the language that suits your query and tools. Language support and operation coverage are documented separately; a shared graph does not make every language feature identical.
SPARQL, Cypher, and GQL queries benefit from the same reasoning capabilities. Deductive AI applies ontologies and rules to the shared graph, making inferred facts available to every supported query language.
Find the people connected to an agreement, the documents behind a decision, or the relationships that satisfy a business rule.
Properties and RDF literals can hold whole files, including documents and recordings.
Use full-text search for words and phrases, or similarity search to find related material before following graph connections.
System metadata is also queryable. Inspect the descriptions and context exposed by the workspace instead of maintaining a separate, disconnected inventory.
Authorized query delegation lets a capable peer do work on behalf of another node. The selected workspace, branch, and access context remain part of that request.
Delegating execution within one logical dataset differs from querying several independent datasets. The querying guide describes the supported query surfaces and examples.
Developer reference
SPARQL 1.1 operates over RDF datasets: subject/predicate/object statements partitioned into named graphs. Cypher and ISO GQL expose vertices, edges and properties through the labelled property graph view.
The execution contract binds language, dataset, workspace, branch/commit, parameters and acting authority. Inferred relationships depend on the workspace's rule profile. A SPARQL basic graph pattern such as ?agreement ex:party ?person and a property-graph traversal can express corresponding relationships without sharing syntax or datatype rules.
Language support is defined per construct and operation. Shared storage does not imply identical grammar, mutation coverage or error behavior. Cross-workspace evaluation requires explicit authority for every participating scope.