Datasets:

geomelon
/

cities

license: cc0-1.0
pretty_name: Geomelon — World Cities, Regions & Countries
short_description: >-
  Multilingual city, region, and country data sourced from Wikidata, one config
  per country, CC0 public domain.
language:
  - multilingual
tags:
  - geography
  - geospatial
  - cities
  - gazetteer
  - places
  - ner
  - rag
  - wikidata
task_categories:
  - text-classification
  - token-classification
size_categories:
  - 100K<n<1M
configs:
  - config_name: spain
    data_files: data/spain/cities.jsonl
  - config_name: united-states
    data_files: data/united-states/cities.jsonl
  - config_name: india
    data_files: data/india/cities.jsonl
  - config_name: united-kingdom
    data_files: data/united-kingdom/cities.jsonl

Geomelon — World Cities, Regions & Countries

Multilingual geographic reference data — cities, regions, and countries — sourced from Wikidata and maintained by Geomelon. Every city carries population, coordinates, elevation, area, postal/dialing codes, settlement type, and name translations into 50+ languages. Released under CC0 1.0 — public domain, no attribution required, free for commercial use.

One config per country (see the dropdown above) — pick a country to avoid downloading the whole corpus. Coverage grows as more countries are exported; see the full current list in the configs block or at geomelon.dev/data-packages.

from datasets import load_dataset

spain = load_dataset("geomelon/cities", "spain", split="train")
spain[0]
# {'id': '...', 'name': 'Barcelona', 'translations': {'ca': 'Barcelona', 'ja': 'バルセロナ', ...},
#  'population': 1620343, 'latitude': 41.3825, 'longitude': 2.1769, ...}

This is one of several ways to use Geomelon data

This dataset is a static, offline-friendly export. If you need something else:

Need	Use
Live, always-current, searchable API	Geomelon API on RapidAPI — free tier, no credit card
Free keyless city autocomplete (no API key)	Free City Autocomplete API
Data shipped with your JS/TS build instead of an npm dependency on `datasets`	Per-country npm packages — `@geomelon/<country>`, same source, same license
Typed TypeScript/JS client	`geomelon` on npm — zero-dependency
Typed Python client	`geomelon` on PyPI — zero-dependency, 3.8+
React autocomplete component	`geomelon-react` — headless hook + component
AI agent / LLM tool-calling access	MCP server (`geomelon-mcp`) — also hosted keylessly at `https://mcp.geomelon.dev/mcp`
No-code / workflow automation	n8n community node (`n8n-nodes-geomelon`)
Full docs, OpenAPI spec, everything above in one place	geomelon.dev

Fields

Country: name, ISO code, capital(s), head of state/government, telephone code(s), driving side, translations.
Region: name, population, coordinates, translations.
City: name, population, coordinates, elevation, area, postal code(s), dialing code(s), settlement type(s), time zone, official website, translations (top-50 world languages plus the country's own official language(s)).

Source & license

Derived from Wikidata (https://www.wikidata.org) via SPARQL. Released CC0 1.0 — public domain, same as the npm data packages and the live API's underlying data. Wikidata data is provided "as is"; not affiliated with or endorsed by the Wikimedia Foundation. A link back to geomelon.dev is appreciated but never required.

Updates

united-kingdom v1.0.0 — 2026-09-07 — Initial publish: 13082 cities, 12 regions.
india v1.0.0 — 2026-08-31 — Initial publish: 50000 cities, 35 regions.
united-states v1.0.0 — 2026-08-29 — Initial publish: 46565 cities, 56 regions.
spain v1.0.2 — 2026-08-29 — +4596 cities, +2 regions.
spain v1.0.1 — 2026-08-09 — README: add FAQ section.
spain v1.0.0 — 2026-08-09 — Initial publish: 19423 cities, 17 regions.