Getting Started
GMEOW is a large ontology, but adoption does not start with the whole graph. Start with one slice, one example, and one projection target. You do not need a reasoner, Docker, Java, or an RDF editor to inspect the vocabulary.
Install
pip install gmeow
Export the bundled docs
gmeow export-docs --directory gmeow-docs
The exported directory is the same static documentation bundle as this site: Markdown, HTML, SVG diagrams, slice pages, term pages, linkage tables, and RDF 1.2 statement documentation.
Pick a first path
For fuller walkthroughs, start with Learning Paths or Recipes. Each recipe points to canonical Turtle examples and the term pages that explain the pattern. The generated Examples catalog lists every slice-local Turtle example when you want broader coverage.
| If you are modelling... | Start with | Then inspect |
|---|---|---|
| people, names, pronouns, aliases, and display suppression | slices/core/names/examples/person-names.ttl |
gmeow:PersonName, gmeow:NameUsage, gmeow:displayable |
| contested or attributed facts | slices/core/standpoint/examples/contested-authorship.ttl |
gmeow:StandpointClaim, gmeow:vantage, gmeow:claimModality |
| documents and public web presence | slices/core/documents/examples/web-presence.ttl |
gmeow:Document, gmeow:webUrl, schema.org linkages |
| events and participants | slices/core/events/examples/wedding.ttl |
gmeow:Event, gmeow:Participation, temporal frames |
| time intervals and calendars | slices/core/temporal/examples/intervals-and-frames.ttl |
gmeow:TimeInterval, gmeow:TemporalFrame |
| creative works | slices/core/creative-works/examples/wemi-novel.ttl |
WEMI, titles, releases, and external work alignments |
| offline distribution and transport | slices/core/gts/examples/dist-package.ttl |
docs/GTS-SPEC.md, gmeow:GTSProfile, gmeow:GTSSegment |
| agent memory or tool trajectories | slices/extensions/agentic/examples/agent-trajectory.ttl |
agentic and provenance terms |
| graph-RAG datasets and pipelines | slices/extensions/graphrag/examples/lillith-dataset.ttl |
dataset, source, chunk, and extraction terms |
Inspect terms while reading examples
gmeow describe gmeow:Person
gmeow describe gmeow:NameUsage
gmeow describe gmeow:StandpointClaim
Every generated term page answers four questions: what the term means, which slice owns it, how it links to other GMEOW terms, and how it projects to external vocabularies.
Read slices as doctrine, not just reference
A slice page explains why a modelling pattern exists. The term pages then give exact class/property details. Useful first slices:
- names: co-equal names, pronouns, usage contexts, and display suppression.
- standpoint: claims that coexist by vantage instead of collapsing to one truth slot.
- temporal: frame-relative time values and solver boundaries.
- teleology: goals, desires, intentions, commitments, and why
gufo:IntrinsicModeappears. - gts: offline transport, profiles, segments, codecs, and compaction lineage.
Follow external links deliberately
When the docs mention PROV-O, P-Plan, gUFO, FIBO, CIDOC CRM, Wikidata, or ConceptNet/ATOMIC, those links point to the generated External Ontologies catalog first. That page lists the target, license, description, and upstream website/namespace. Formal external terms such as gufo:IntrinsicMode point to External Terms, where the docs translate the term before sending you to the upstream ontology.
Understand statement metadata
GMEOW uses RDF 1.2 / RDF-star style statement metadata for provenance, confidence, temporal scope, and standpoint. If a fact needs an accordingTo, confidence, validity interval, or attribution, read the RDF 1.2 statement layer before flattening it.
Use linkages as adoption maps
The Linkages page is generated from the mapping DSL. It shows SSSOM alignments, projection profiles, lossy-drop notes, and external vocabulary coverage. Treat it as the map from native GMEOW to consumer formats such as schema.org, PROV-O, vCard, FOAF, GeoSPARQL, and Wikidata.
Use the web docs and content negotiation
Human requests for GMEOW IRIs resolve into this static site. RDF clients continue to use HTTP content negotiation for Turtle, RDF/XML, N-Triples, and JSON-LD serializations.