AI Localization × Human Experts

Global and public-sector relationships

Alongside our professional-service clients, LETR WORKS has built relationships with global businesses, professional-service companies, and public institutions.

Sector	Selected clients and partners
Global companies and professional services	Kantar, Concentrix, Pengtai, Trellis, Morningside, Elephant Around
Public institutions	NIA, NIPA, K-DATA, Korea Tourism Organization, TTA, Korean Standards Association, KOITA, KOICA
Media	AVING News

These organizations are listed in the official LETR WORKS client and partner introduction.

Selected projects

Jecheon Cultural Foundation — multilingual cultural videos

LETR WORKS supported Vietnamese, Russian, Chinese, and English subtitles for the foundation’s 2025 “Culture and Arts RO” YouTube project. Foreign residents associated with local multicultural-family and overseas-Korean support centers participated in translation and review, using AI video translation to make local cultural content accessible in more languages.

Read the Jecheon Cultural Foundation case

Webtoon studio — translation, typesetting, and editable delivery

An operational workflow connects dialogue extraction from PSD files with translation, typesetting, and layer reconstruction. Editable PSD files and review images allow the studio to continue post-editing in its own production process.

Global pharmaceutical company — specialist document translation

An operational document workflow handles PDF, Word, and HWP files with customer-specific terminology and delivery to the company’s FTP server. Terminology management, document formats, and delivery are connected in one process.

Video post-production and OTT delivery — subtitle and output checks

Post-production workflows support MXF track checks, platform-specific subtitle specifications, subtitle burn-in, and large-file sharing. OTT delivery workflows connect speaker audio tracks, audio processing, and outputs in multiple formats.

Broadcasting and OTT technology validation

A 2023 public account described video translation benchmarking in progress with SBS and Content Wavve, alongside PoCs and usage contracts with global companies and Korean translation agencies.

Read the broadcasting and OTT validation account (Korean)

Deliverables by content type

Video, webtoons, and specialist documents differ in work units and delivery specifications as well as language. AI prepares drafts; translators and reviewers check expression, context, and formatting before delivery.

Content	Work scope	Main outputs
Video, broadcast, and OTT	Dialogue extraction, subtitle translation, dubbing, timestamped review	Translated subtitles, dubbed audio, and video in agreed specifications
Webtoons and digital comics	Dialogue translation and text-layer layout	Editable PSD and review PNG files
Specialist documents	Terminology, source comparison, tables and footnotes	Translated documents in the agreed format and review results

Webtoons: translation and typesetting together

Editable PSD files allow direct extraction from text layers. Text merged into artwork needs a different path such as OCR, so the method depends on the original. In the text-layer path, translated text is placed without regenerating the background.

Layout takes account of speech bubbles, dialogue order, fonts, and vertical writing. Character voice, sound effects, and proper names are reviewed with glossaries and experts. Editable files and review outputs are prepared together.

Video: subtitles, dubbing, and timestamped review

Inputs include MP4, MOV, MKV, and MXF video and SRT, VTT, and TTML subtitles. Speech recognition and speaker separation support the draft. Translators and reviewers compare expression and timing with the original segment.

Reviewers and project managers can leave timestamped comments, approvals, and revision requests. Subtitle specifications, broadcast track checks, burn-in rendering, and format conversion are scoped against the source and delivery requirements. Transfer paths such as FTP or YouTube are confirmed per project.

Documents: terminology and structure

PDF, DOCX, PPTX, XLSX, HWPX, and XML documents are processed with their text and structure. Scans require layout analysis and OCR. Tables, footnotes, and formatting are reviewed against the source, using existing translation assets for domains such as patents, public administration, and technical documents.

Reuse approved translations

Approved translations accumulate in customer-specific translation memory (TM); names and specialist terms are kept in glossaries (TB). Repeated sentences and terms can refer back to approved wording. This supports continuity across episodes, series, and recurring documents.

Existing TM and glossaries can be imported. Customer content is not used to train shared or general-purpose models. Dedicated environments and processing scope are agreed against security needs.

Service and operating options

Option	Suitable situation	Scope to confirm
Managed expert service	Translators and reviewers are needed	Languages, expertise, volume, quality criteria, and deadline
Team-operated platform	Your own translation and review team	LETR WORKS features and operating configuration
System integration	Connection to an existing CMS or DAM	APIs, file transfer, and operational responsibilities
Dedicated deployment	Separate security requirements	Environment, data handling, and operations

Source files and delivery requirements are reviewed first. Sample work establishes quality, formatting, and turnaround. See the platform scope in LETR WORKS.