About the XEO project
xeo: · v0.5.0 · open draft · six layers · eight modules · CC BY 4.0
A small, standards-grounded vocabulary for how engines — search, generative, answer, conversational — see, cite, and serve web content. One graph model for the whole xEO family.
Why "XEO"
Marketers already think in -EO: SEO, then GEO, then AEO. The engines keep changing; the discipline — make your content legible and creditable to the machine between you and your audience — doesn't. XEO names that discipline once, and the ontology's Engine taxonomy absorbs new surfaces as subclasses (SearchEngine, GenerativeEngine, AnswerEngine, ConversationalAgent) instead of new acronyms.
The x is the variable. That is the whole point of the name: the letter in front keeps changing, and the discipline behind it doesn't.
Status & versioning
XEO is at v0.5.0 — an open draft, built to be riffed on. Issues and pull requests are welcome.
Maturity varies by module, and each one says so:
xeo-core— the stable substrate the rest of the ontology builds on.xeo-visibility— field-tested against a working tracker / response-capture / Neo4j implementation.xeo-engagement— an early draft published for discussion: the "pre-click funnel" seam that engine-side data has been missing.xeo-decisionandxeo-strategy— implemented in production (seexeo-graph-kitfor a runnable slice).xeo-schemes— every value lifted verbatim from that production implementation.
Six epistemic layers type every term by the kind of claim it makes — entity, episode, observation, derivation, interpretation, norm — and eight modules cut the vocabulary by concern.
Namespace
Terms resolve at xeoontology.org, not at the repo. Two hash namespaces, and nothing else:
xeo:— https://xeoontology.org/xeo# — every term lands on /xeo (/xeo#Gap,/xeo#addressableBy, …).xeos:— https://xeoontology.org/schemes# — every scheme and concept lands on /schemes (/schemes#gapType-missing_page).
Standards & grounding
XEO invents no semantics where established ones exist. Observations are grounded in primary standards, and xeo-align.ttl is an optional crosswalk — alignments, not dependencies:
- HTTP — RFC 9110 (semantics, status codes, redirection) grounds fetched, statusCode, redirectsTo.
- RFC 6596 — the canonical link relation grounds hasCanonical.
- WHATWG HTML / URL — ground linksTo, title, metaDescription, address, origin.
- The Google Search Console Search Analytics API — grounds clicks, impressions, ctr, avgPosition and the ranksFor rollups.
- Broder (2002), "A taxonomy of web search" — the intent taxonomy behind the four SearchIntent individuals, extended in common practice.
- schema.org, PROV-O, SKOS — the neutral entity base, the provenance spine, and the taxonomy spine, bridged in
xeo-align.ttlwith skos:closeMatch where the correspondence is approximate.
Prior art: SEOntology (seovoc) — the closest one, an open-source SEO domain ontology by WordLift et al., accepted at SEMANTiCS 2026. XEO cites it and expects to align where the vocabularies genuinely meet; the decision-intelligence and assistant-surface layers are XEO's reason to exist. See the prior-art notes on the home page.
Machine-readable files
Everything machine-readable is served from this site root, byte-identical to the repo:
xeo-core.ttl·xeo-visibility.ttl·xeo-engagement.ttl·xeo-decision.ttl·xeo-strategy.ttl·xeo-delivery.ttl·xeo-schemes.ttl·xeo-align.ttl— plain OWL / SKOS; load into any triple store or ontology editor. Decision builds on core; strategy and delivery build on decision; the rest stand alone.context.jsonld— a JSON-LD@contextmapping graph labels/relationships/properties to IRIs: point it at a Neo4j export and it validates as RDF/JSON-LD.schemes.json— a generated projection ofxeo-schemes.ttlfor anything that builds an enum at boot.schema.cypher— Neo4j 5.x constraints and indexes for every module.
The same files are linked from the /xeo term reference header, or browse them in the repo.
Maintained by — and license
XEO is maintained by Inbound Found, built by working backwards from a production marketing knowledge graph, then generalized. Contributions, counter-proposals, and rude questions about our modeling choices are all welcome.
License: CC BY 4.0 — use it, extend it, ship it; just attribute.