VL Viktor Lavrov

Web app · open data

Languages of the World

An interactive atlas of 7,992 languages and 13,706 dialects: who speaks them, where, how endangered they are — and how they relate to each other, in four different senses.

  • Free
  • Open source
  • No account
  • 15 interface languages
Open the atlas Source on GitHub
The atlas: a sortable table of languages beside a world map where every language is a dot coloured by endangerment status. The atlas: a sortable table of languages beside a world map where every language is a dot coloured by endangerment status. Shown in the app's dark theme.

What it looks like

A language card for Russian with family, macroarea, countries, speaker count, alternative names and a list of dialects. A language card for Russian with family, macroarea, countries, speaker count, alternative names and a list of dialects. Shown in the app's dark theme.
A card per language — family, countries, speakers, dialects, links
The Language connections panel listing the Glottolog branch, closest relatives, languages with similar grammar and vocabulary as percentages, and neighbours on the map with distances. The Language connections panel listing the Glottolog branch, closest relatives, languages with similar grammar and vocabulary as percentages, and neighbours on the map with distances. Shown in the app's dark theme.
Four kinds of connection, each measured separately
A side-by-side comparison of Russian and Ukrainian: basic parameters, then WALS grammar features with matching rows highlighted, reading 53 of 56 shared features match. A side-by-side comparison of Russian and Ukrainian: basic parameters, then WALS grammar features with matching rows highlighted, reading 53 of 56 shared features match. Shown in the app's dark theme.
Compare any two languages, feature by feature

What it does

Where the data comes from

Four open datasets, all credited in the footer of the atlas itself: Glottolog for languages, dialects, families, coordinates and endangerment status; Wikidata for speaker counts and localised names; WALS for the 192 grammar features; and ASJP for the Swadesh lists behind lexical similarity.

It's a static site — the data ships as JSON and everything runs in your browser. No account, no tracking, and the heavier files (alternative names, WALS) load only when something needs them.

Worth knowing: the numbers are aggregated automatically. Speaker counts exist for roughly 1,800 languages and can be out of date for small ones, and similarity scores are a measurement, not a verdict — similar words can come from borrowing rather than kinship. Check the primary sources before citing anything.

Running it yourself

The whole site is the web/ folder of the repository — any static host will do. Locally it needs a server only so the page can fetch its JSON:

Clone it on GitHub

The data/ folder holds the Python scripts that rebuild everything from the upstream sources, so the atlas can be refreshed as they publish new releases.