LangDex

Localization & Translation

Professional translators spend 30% of their time looking up terminology. LangDex's Analyze lens delivers register metadata (sonkeigo vs casual), domain classification, collocations, and full morphological paradigms—directly in your CAT tool, no context-switching required.

The Challenge

The Terminology Lookup Tax

  • 1Translators alt-tab to dictionaries 50+ times per hour—each lookup breaks flow and costs seconds that compound
  • 2Static termbases don't encode register: is this term honorific (敬語), humble (謙譲語), or casual? Critical for Japanese/Korean
  • 3MT hallucinates terminology for domain-specific content (legal, medical). No automated verification layer exists
  • 4Japanese orthographic variation (漢字/ひらがな/カタカナ/romaji) means searches miss synonymous forms
  • 5QA is manual: spotting mistranslated named entities or missing terms requires human review of every segment
The Solution

In-Editor Terminology Intelligence

LangDex's Analyze lens projects terminology data optimized for professional translation. Select a term, receive: definition, register tags (formal/casual/honorific/humble), domain classification (legal/medical/technical), typical collocations from corpus data, and full UniMorph paradigms for morphologically rich languages. All via a single API call that fits CAT tool response budgets.

Features

Key Capabilities

Register Metadata

Japanese politeness levels (敬語/謙譲語/丁寧語), Korean speech levels, formality markers. Essential for business and literary translation.

Domain Classification

Sense-level domain tags: legal, medical, technical, colloquial. Disambiguate polysemous terms by context.

Collocation Patterns

Corpus-derived collocates: which verbs pair with this noun? Which particles? Write target text that sounds native.

CAT Tool Integration

REST API designed for Trados, MemoQ, Phrase TMS latency requirements. Sub-100ms terminology lookups.

Named Entity Alignment

Compare source and target for entity consistency. Flag missing company names, product terms, or mistranslated proper nouns.

Orthographic Normalization

Query '東京' or 'とうきょう' or 'Tokyo'—resolve to the same lexeme. Full CJK reading/writing variant mapping.

Who Is This For

Perfect For

Translation agencies (LSPs)In-house localization teamsCAT tool developers (Trados, MemoQ, Phrase)Language service providersGame localization studios (JRPG, visual novels)Legal/medical translation specialistsTechnical documentation teamsTerminology managers building termbases

Ready to Get Started?

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