TransLens

It reads the Korean or English source and surfaces the terminology. You do the translating.

Open Run locally

Free and open source. Analysis only — no machine translation.

Source analysis

L1 lexical flags, L2 Hanja and morphology, L3 tradition frames. The tool never writes the target.

Project glossary

Import CSV or JSON, keep a locked kit on the local workbench, and match terms as you analyze.

Legal reference

Serve your own topic-filtered reference locally. Proprietary corpora stay off the public site.

Three levels you dial in

Contextual choices are rule-pack driven, not generic disambiguation.

  • L1 · LexicalMisleading calques, category mismatch, and public-domain Webster sense hints.
  • L2 · EtymologyHanja decomposition, Sino-Korean morphology, and per-segment glossary hits.
  • L3 · TraditionFrame mismatch across Buddhist, Taoist, and Confucian readings. Skeleton today — L2 still runs.
  • No machine translationThe tool analyzes and references. You, or a tool you choose, write the target.
  • Local by defaultSelf-hosted and offline-capable. DualDraft stays a separate drafting app.
  • Project-scoped dataGlossary and reference stay filtered per project without an explicit share path.

Run locally

Install once, then open the workbench in your browser. The full glossary and legal reference stay on your machine.

Open the repository

The public Open path is a read-only workbench. Writes stay local.

Setup commands · macOS / Linux 
git clone https://codeberg.org/russkysong/translens.git
cd translens
pnpm install
pnpm setup
pnpm dev

Open http://127.0.0.1:5173 after the servers start.

Languages
Korean and English legal translation
Processing
Source analysis on your machine. No built-in machine translation.
Requirements
Node.js 22 and pnpm 9
Other uses
Codeberg