Linguistic Infrastructure for the World
We're building the semantic layer that connects 5,700+ languages. Enabling AI systems, language learners, and researchers to traverse the full depth of human linguistic knowledge.
Language Data is Broken
For decades, digital language tools have relied on legacy systems that fail to capture the richness of human language.
Flat-File Dictionaries
Legacy databases like JMdict and CC-CEDICT store words as isolated entries. No semantic connections, no etymological links, no context.
Context Poverty
Traditional dictionaries can't distinguish homophones or explain why a word fits a specific context. Apps become indistinguishable 'dictionary wrappers.'
The AI Language Gap
Modern AI concentrates on a handful of high-resource languages. The other 7,000+ languages remain invisible to algorithms driving the global economy.
Language as a Graph
LangDex models language as an interconnected network of entities and relationships. Not a flat list of definitions.
Graph Architecture
Words, kanji, radicals, and historical variants exist as nodes. Etymology, semantics, and morphology form the edges that connect them.
The Concept Entity
Our integration with PanLex provides language-agnostic Concept IDs. Stable semantic anchors that connect meanings across 5,700+ languages.
The Lens System
Learn, Analyze, and Research lenses filter the graph for specific users. From beginners to professional translators to academic researchers.
Comprehensive Coverage
We've aggregated and structured the world's best open linguistic datasets into a unified, queryable graph.
5,700+
Languages
15M+
Lexemes
100M+
Word Forms
2M+
Cross-lingual Concepts
Built on Open Data
LangDex integrates data from PanLex, Glottolog, Wiktionary, JMDict, KanjiDic2, UniMorph, Tatoeba, and more. Structured into a high-fidelity linguistic graph.
Bridge the Language Divide
We believe linguistic data should be structured, interconnected, and accessible. LangDex exists to transform fragmented dictionary files into the semantic infrastructure that powers the next generation of language technology. From AI systems that don't hallucinate, to learning apps that understand context, to humanitarian tools that work on Day Zero of any crisis.
Accessibility
Language knowledge shouldn't be locked behind paywalls or trapped in incompatible formats. We're building infrastructure for everyone.
Innovation
From Neuro-Symbolic AI to GraphRAG, we're positioned at the intersection of linguistic scholarship and cutting-edge technology.
Interoperability
Our Concept IDs create a universal 'foreign key' that links languages across AI models, databases, and applications.