dotNetRDF

dotNetRDF is an open-source library that provides a complete API for parsing, managing, querying, and writing RDF (the Resource Description Framework) in .NET environments. Released under the permissive MIT license and distributed via NuGet, it is the de facto RDF toolkit for the C#/.NET ecosystem, filling the role that tool:rdflib fills for Python or tool:apache-jena fills for Java. The project was authored by Rob Vesse and a small group of collaborators (including Ron Michael Zettlemoyer, Khalil Ahmed, Graham Moore, Tomasz Pluskiewicz, and Samu Lang) and is maintained as a community effort on GitHub.

The library models RDF as an in-memory graph and triple/quad store, exposing typed nodes (IRIs, blank nodes, and literals with datatypes and language tags), triples, and named graphs. Around this core it ships reader and writer implementations for the standard RDF serializations, including format:turtle, format:n-triples, format:rdf-nquads, format:trig, format:trix, format:rdf-xml, format:rdf-json, format:notation3, and format:json-ld. Because these formats are interchangeable views over the same triple model, dotNetRDF effectively acts as a format-conversion hub: parse one serialization into a graph and write it back out as any other. As of the 3.x release it also supports format:rdf-star and the corresponding SPARQL-Star extensions.

Querying and updating are handled by dotNetRDF's own Leviathan engine, which implements SPARQL 1.1 Query and Update. Optional modules extend this with full-text search over the Leviathan engine, SPIN, basic inferencing (RDF Schema, SKOS, and a subset of OWL, related to format:owl), and SHACL validation. A client package provides a common abstraction over external triple stores such as AllegroGraph, Stardog, Virtuoso, and Jena-backed endpoints, and an ASP.NET package can host SPARQL Protocol and Graph Store Protocol services.

Version 3.x targets .NET Standard 2.0 (covering .NET Framework 4.7+, .NET Core 2.0+, and .NET 5+) and was restructured into smaller packages to minimize dependencies and separate the core from higher-level APIs. Strengths include broad standards coverage, a mature SPARQL implementation, and native integration for .NET applications. Limitations are that it is scoped to the RDF/semantic-web stack rather than general labeled-property-graph tooling, the in-memory model constrains very large datasets to external stores, and the 3.x rewrite introduced breaking API changes from earlier 2.x releases.