Academic Research

Cross-linguistic research requires assembling datasets from PHOIBLE, UniMorph, Glottolog, WALS, and Wiktionary—each with different schemas, formats, and access methods. LangDex unifies these into a single queryable graph: phoneme inventories, morphological paradigms, genealogical relationships, and etymological chains accessible via one API.

The Challenge

The Data Fragmentation Problem

1PHOIBLE, UniMorph, Glottolog, WALS each have different schemas—cross-referencing requires custom ETL per project
2Endangered language data scattered across PDFs, Toolbox files, FLEx databases, and researchers' hard drives
3IPA transcription inconsistencies across sources: is it /ɾ/ or /r/? Normalization is manual
4Assembling a typology dataset (e.g., 'all languages with ejective consonants') takes weeks of data wrangling
5Reproducibility fails: 'I queried PHOIBLE in 2019' doesn't specify version, access method, or filtering

The Solution

Unified Research Graph

LangDex's Research lens exposes the full depth of integrated linguistic datasets. Query PHOIBLE's 3K+ phoneme inventories, traverse UniMorph's morphological paradigms for 150+ languages, filter by Glottolog genealogy (27K languoids), and trace Wiktionary etymology chains—all in one API call. Export to CLDF for interoperability. Cite with version-pinned URIs for reproducibility.

Features

Key Capabilities

PHOIBLE Integration

3,020 phoneme inventories from 2,186 languages. Query: 'all languages with /ɬ/' or 'Sino-Tibetan tone systems.' Standardized IPA.

UniMorph Paradigms

Morphological inflection tables for 150+ languages. Query full verbal paradigms, nominal declensions, comparative forms.

Glottolog Genealogy

27K languoids in hierarchical tree. Filter by family (Austronesian), genus (Oceanic), or macroarea (Papunesia).

Etymology Graphs

Trace cognate sets across language families. Identify sound correspondences. Link to Proto-forms where documented.

CLDF Export

Export datasets in Cross-Linguistic Data Format. Compatible with cldfbench, lingpy, and standard comparative tools.

Versioned Citations

Every query returns version-pinned URIs. Cite exact dataset state for reproducible research.

Who Is This For

Perfect For

Typology researchers (WALS contributors)Computational/corpus linguistsHistorical/comparative linguisticsPhoneticians and phonologistsNLP researchers (multilingual models)PhD students assembling dissertation corporaLanguage documentation projectsPsycholinguistics (cross-linguistic stimuli)

Ready to Get Started?

Join our waitlist to get early access to LangDex's API and start building.