Open data

Everything on this site is a JSON file.

The site is static and every page renders from the files below. Use them directly, mirror them, or rebuild them from the upstream sources with one script.

Last built 2026-10-04 · 7,567 languages · 175 scripts

Files

JSON

Language index

data/languages/index.json

One compact row per language: ISO 639-3, name, macroarea, family, endangerment, scripts, coordinates, speakers and GlotSuite coverage flags.

i: ISO 639-3
n: name
m: macroarea
f: family
e: endangerment level (1–6)
s: scripts (ISO 15924)
t: coverage flags
y: latitude
x: longitude
p: speakers

Download

JSON

Language details

data/languages/details/{letter}.json

Booklet data sharded by the first letter of the ISO code: endonym, countries, Glottocode, Wikidata ID, documentation level, auxiliary scripts, GlotLID per-label metrics and Glot500 labels.

Download (a.json)

JSON

Scripts

data/scripts.json

One record per ISO 15924 script: name, type, direction, Unicode ranges, specimen, languages using it, GlotLID label count and per-model OCR results.

Download

JSON

OCR models

data/ocr_models.json

GlotOCR Bench summary per model: overall, high/mid/low-tier accuracy and macro CER.

Download

JSON

Projects

data/projects.json

GlotSuite projects with descriptions, links, usage snippets and BibTeX.

Download

JSON

Statistics

data/generated_stats.json

Counts derived when the data was last built.

Download

JSON

Site copy

data/site.json

All page text, navigation and team information.

Download

JSON

Vocabularies

data/taxonomy.json

Labels and colours for macroareas, endangerment levels, coverage flags and script types.

Download

JSON

Legal pages

data/legal.json

Text of the Privacy Policy and Terms of Use.

Download

Rebuild from source

git clone --depth 1 https://github.com/glottolog/glottolog-cldf src/glottolog-cldf
for r in GlotScript GlotLID GlotWeb GlotOCR-bench; do
  git clone --depth 1 https://github.com/cisnlp/$r src/$r
done
pip install pycountry
python3 tools/build_data.py --sources src

Sources & licences

Source	Used for	Licence
Glottolog	Names, Glottocodes, families, macroareas, coordinates, endangerment, documentation level	CC BY 4.0
GlotScript-R	Writing systems per language, Unicode ranges per script	MIT
GlotLID v3	Supported labels and per-label F1 / precision / recall	Apache-2.0
LinguaMeta (via GlotWeb)	Endonyms, speaker estimates, Wikidata IDs	CC BY 4.0
Glot500 label list (via GlotWeb)	Languages covered by Glot500	Apache-2.0
GlotOCR Bench	Per-script OCR accuracy and CER for 14 models	see repository