VisualArtsDNA began as a simple question: How can the relationships within a body of artwork be made visible, explicit, and shareable? What started as a personal experiment in organizing creative work has grown into a comprehensive information model for the visual arts, supported by modern semantic-web technologies — and now, an open source platform available to any artist who wants to host their own catalog.
The project explores how artworks, materials, processes, themes, and artistic intentions connect to one another. By structuring this information, VisualArtsDNA provides a framework for documenting creative practice, analyzing patterns, and supporting discovery. While the home page foregrounds the artworks themselves, this site also serves as the public home for the underlying model that organizes them.
The structured data has a practical benefit beyond the site itself: search engines index the semantic relationships, not just the page text. The correct information about each work — dimensions, materials, date, process — is available to any system that can read linked data.
At the core of VisualArtsDNA is an ontology designed to represent the domain of visual arts with precision and flexibility. Key features include:
The ontology includes:
This structure allows the system to reflect not just what a work is, but how it relates to other works, influences, and artistic choices. Process notes and criticism are linked to the works they concern, creating a record of artistic thinking as well as artistic output.
Every artwork displayed on the site is represented internally as an RDF instance. Together, these instances form a growing knowledge graph of creative works. This graph enables:
This site is served by cwvaServer-py — a Python/FastAPI semantic web server ported from the original Groovy/Jetty implementation. The port was developed collaboratively with Claude (Anthropic's AI assistant) using Claude Code, Anthropic's agentic coding tool.
cwvaServer-py is open source and available to any artist or collective who wants to host their own structured catalog of creative works. It runs on a laptop, a home network, or a free-tier cloud instance. No external database required — data lives in standard text files that are human-readable, version-controllable, and portable. No platform dependency.
For large-scale deployments, Apache Jena Fuseki + TDB2 provides a persistent indexed triplestore backend. See DATABASE.md in the repository for details.
VisualArtsDNA is an evolving project. The ontology continues to expand as new artworks, concepts, and relationships are added. Planned development includes:
See the project roadmap for the full picture.