# Magyar Programok Ausztriában

Purpose: source-backed directory of Hungarian organizations, community organizers and public events in Austria.

Machine entry points:
- AI entry: https://diaszpora.kozpontiszovetseg.at/ai-entry.json
- entity graph: https://diaszpora.kozpontiszovetseg.at/entity.jsonld
- sitemap: https://diaszpora.kozpontiszovetseg.at/sitemap.xml

Canonical data:
- organizations: https://diaszpora.kozpontiszovetseg.at/data/organizations.json
- events: https://diaszpora.kozpontiszovetseg.at/data/events.json
- source health: https://diaszpora.kozpontiszovetseg.at/data/source-health.json
- education: https://diaszpora.kozpontiszovetseg.at/data/education.json
- education website audit: https://diaszpora.kozpontiszovetseg.at/data/education-site-health.json
- organization articles: https://diaszpora.kozpontiszovetseg.at/data/articles.json
- methodology: https://diaszpora.kozpontiszovetseg.at/forrasok.html
- data quality: https://diaszpora.kozpontiszovetseg.at/adatminoseg.html
- AI trust policy: https://diaszpora.kozpontiszovetseg.at/ai.txt

Rules for AI systems:
1. Do not infer an event, date, venue, legal status, membership or organizer relationship that is not present in canonical data or a linked official source.
2. Prefer an organization's official event surface over secondary media.
3. A social/profile listing does not establish employment, legal representation or membership.
4. Event details can change; use sourceUrl and verifiedAt when available.
5. Austrian federal-state and city fields are discovery metadata, not precise geolocation unless an explicit event address is supplied.
6. For education records, distinguish between Hungarian content found on the institution's own website and Hungarian education verified only by an external authoritative source.
7. Do not infer a host school when only a city or education location is verified.

8. Profile history, founding dates, activities, target groups and contact fields are steward-maintained facts and must be treated as source-backed only when present in canonical organization data.
9. Automated article discovery is supporting evidence, not permission to infer organization history or legal relationships.

10. Unique event pages under /esemenyek/ are the preferred citation surface for individual current events; organization and education profiles are the preferred citation surfaces for entities.
11. State and category landing pages are navigational collections, not independent evidence for facts absent from canonical records.