About
Preserving cultural heritage through knowledge graphs
HeritageGraph is a platform by CAIR-Nepal for documenting cultural heritage — with a focus on Nepalese and Himalayan heritage — and publishing it as Linked Open Data modelled in CIDOC-CRM.
Our mission
Cultural heritage is the living memory of communities — traditions, artifacts, stories, and practices that connect the present to the past. Much of it remains undocumented, fragmented across institutions, or at risk of being lost.
HeritageGraph builds an interconnected digital record of that heritage using knowledge graph technologies. Publishing heritage data as Linked Open Data following CIDOC-CRM lets researchers, communities, and institutions trace connections, cite sources, and reuse the material under clear terms.
Core features
Our knowledge graph
The model behind every record: entity types, the relations allowed between them, and the live graph of what has been published so far.
Data, provenance and reuse
What a reader, reviewer, or reuser needs to know before relying on these records.
- Provenance
- Imported sources are kept in frozen, separately named graphs. Curated claims carry a data source and a confidence value, so every published statement can be traced to where it came from.
- Access tiers (CARE)
- Records can be marked public, restricted, community-only, or culturally sensitive. The SPARQL proxy filters restricted material rather than relying on the interface to hide it.
- Licensing
- Software is MIT-licensed. Data carries per-source terms — OpenStreetMap, Wikidata, UNESCO and community contributions differ, and the applicable licence travels with each source.
- Identifiers
- Entities receive persistent w3id.org IRIs. External alignments are recorded as explicit links rather than silent merges.
Standards and vocabularies
The model is not bespoke: it reuses established heritage and semantic web vocabularies so the data is interoperable with other archives.
- CIDOC-CRM
- Core ontology for cultural heritage documentation
- CRMinf
- Argumentation model for beliefs, inference and evidence
- LinkML
- Schema source that generates the ontology and forms
- SHACL
- Shape constraints validating records before publication
- SKOS
- Controlled vocabularies and concept schemes
- VoID / DCAT
- Machine-readable dataset descriptions for discovery
Center for Artificial Intelligence Research Nepal
About CAIR-Nepal
CAIR-Nepal is a non-profit research organization advancing artificial intelligence research and its applications for social good in Nepal and the wider region. HeritageGraph is one of its projects.
Contribute to the record
Documenting heritage is slow, collaborative work. Add a record, cite a source, or review someone else's contribution.
Image credits
- Basantapurpalace.JPG — Artha at English Wikipedia · CC BY-SA 3.0 · Wikimedia Commons
- Patan durbar square.jpg — Zulufive · CC0 · Wikimedia Commons
- View of Bhaktapur Durbar Square.jpg — Bikram Pratap Singh · CC BY-SA 3.0 · Wikimedia Commons
- Swayambhunath 2018.jpg — Nabin K. Sapkota · CC BY-SA 4.0 · Wikimedia Commons
- Boudhanath stupa , Kathmandu, Nepal.jpg — Sumitbhatt222 · CC BY-SA 4.0 · Wikimedia Commons
- Pashupatinath Temple-2020.jpg — Bijay Chaurasia · CC BY-SA 4.0 · Wikimedia Commons
- Nepal - Changu Narayan (3566057331).jpg — Dhilung Kirat from Santa Barbara, CA · CC BY 2.0 · Wikimedia Commons
- BRP Lumbini Mayadevi temple.jpg — Bibek Raj Pandeya · CC BY-SA 4.0 · Wikimedia Commons
All photographs from Wikimedia Commons, retrieved 2026-08-28.
Listing facts on this page come from the UNESCO World Heritage List. Details about each place come from HeritageGraph records a reviewer has checked.
Photography from Wikimedia Commons, retrieved 2026-08-28. Each image carries its licence and author.
Browse everything we hold →Open source HeritageGraph is developed in the open. Issues and contributions are welcome.