June 11, 2026 · 8 min read · Data

1,199 UNESCO World Heritage Sites: where they cluster and where they don't.

The UNESCO World Heritage list is the closest thing travel has to a globally accepted "what's important" register. It's also wildly uneven. Italy and China each have nearly 60 sites; entire sub-regions have fewer than three. If your AI travel product leans on UNESCO as a grounding signal, you need to know where the signal goes silent.

The top of the list

As of the 2025 inscription cycle, the country leaderboard is dominated by a familiar set:

Italy: 60 sites. Rome, Venice, Florence, plus more in Tuscany, Sicily, the Amalfi Coast, the Dolomites.
China: 59 sites. The Great Wall, the Forbidden City, Mount Tai, Huangshan, Lijiang, plus rapid additions in the last decade.
Germany: 54 sites. Cologne Cathedral, Wartburg, the Bauhaus sites, plus a quietly accumulating set of cultural landscapes.
France: 53 sites. Versailles, Mont Saint-Michel, the Loire valley, plus Caribbean and Pacific overseas territories.
Spain: 50 sites. Alhambra, Sagrada Família, Toledo, Santiago de Compostela, multiple Camino routes.
India: 43 sites — appearing in passing here as part of the broader pattern.
Mexico: 35 sites. The most in the Americas, dominated by Mesoamerican archaeological zones plus colonial city centers.
UK, Iran, Russia, Japan: 25-35 each.

The pattern is clear: Mediterranean Europe and Eastern Asia have been on the inscription list since the 1970s and have deep state-level archaeological infrastructure to push nominations forward. The rest of the world is still catching up.

Where the list goes quiet

Whole regions are dramatically underrepresented:

Sub-Saharan Africa: 100 sites total across 47 countries. Many countries have one or zero. The Sahel and Central African states are particularly sparse.
Pacific island nations: A handful of sites total. Vanuatu, Solomon Islands, Tuvalu, Nauru, Niue have zero or one each.
Central Asia: Each Stan has a representative listing or two — the Silk Road serial sites, Itchan Kala in Uzbekistan, Saryarka in Kazakhstan — but nothing approaching the density of comparably sized European regions.
Caribbean: Despite the cultural depth and tourism volume, few sites: Old Havana, Brimstone Hill, Vieille Saint-Pierre, a couple more.
Gulf states: UAE has one (Al Ain). Qatar, Bahrain, Kuwait, Oman have one to five each despite significant heritage tourism investment.

What this means for AI grounding

UNESCO inscription is a strong cultural-anchor signal where it exists. A travel AI told "ground every itinerary item to UNESCO when possible" produces excellent itineraries in Italy (you'd have to try to plan a bad week in Tuscany) and in cities like Cairo, Istanbul, Kyoto where the inscriptions track tourist intent closely.

In sparse regions the same instruction degrades. Asking a model to ground a Cape Town itinerary to UNESCO and the answer is mostly Robben Island and Cape Floral Region. Real Cape Town tourism is the V&A Waterfront, Table Mountain, Boulders Beach penguins, the wine routes — none of which are inscribed.

The pattern: in dense regions, UNESCO is sufficient; in sparse regions, you need other anchors — Foursquare quality tier, OSM tourist tags, editorial circuits, national park systems, museum networks.

The /v1/unesco endpoint

GET /v1/unesco?country=GR

[
  { "name": "Acropolis, Athens",          "year": 1987, "category": "cultural" },
  { "name": "Meteora",                    "year": 1988, "category": "mixed" },
  { "name": "Mount Athos",                "year": 1988, "category": "mixed" },
  { "name": "Old Town of Corfu",          "year": 2007, "category": "cultural" },
  { "name": "Pythagoreion & Heraion of Samos", "year": 1992, "category": "cultural" },
  ...
]

Filterable by country, category (cultural / natural / mixed), and inscription year range. Each row carries the canonical UNESCO ID, so you can cross-reference to the official register without scraping.

The political layer

Inscription is also political. The list reflects which governments have invested in nomination infrastructure, which have stable enough heritage agencies to maintain conservation plans, and which are seen as politically aligned with the voting bloc at any given session. Countries like Iran (29 sites) score above what a casual observer might predict; some smaller European states score above their cultural footprint because they've been working the nomination cycle since the 1970s.

None of which makes UNESCO wrong as a grounding signal. It just means: use it as a strong "definitely worth seeing" hint where it fires, and don't treat its silence as evidence of absence everywhere else.