Localization Technology Management at GitLab

Comprehensive overview of GitLab’s localization technology ecosystem, workflows, and AI-powered translation infrastructure enabling global content delivery.

Mission and vision

Localization technology at GitLab - mission and vision

Localization technology stack - Overview

The Localization team manages a sophisticated technology ecosystem designed to automate and enhance translation workflows across GitLab’s global content. Our technology stack consists of purpose-built custom solutions, commercial Language Technology Platforms (LTPs), and emerging AI-powered services that collectively enable localization of GitLab’s product UI, marketing content, and product documentation.

Localization technology stack - Components

See more details and visuals in the Localization management technology stack at GitLab issue.

Tools inventory

Commercial Tools:

  • Argo as a platform, aka Request Management System and its components (by Spartan Software)
  • Phrase TMS (by Phrase)
  • TranslationOS (TOS) and Matecat (by Translated)
  • Crowdin (by Crowdin)

At GitLab:

  • Google Cloud Platform sandbox with ai-machine-transl project
  • Anthropic Claude UI with emerging standalone projects

Orchestration platform

Argo - Localization Request Management system serving as the central orchestration hub, consisting of the following specialized services:

  • Argo web client (UI) - Argo web UI used by localization program managers, stakeholders and vendors
  • Argo web services - backend/API orchestration engine
  • Argo-Phrase integration - service for GitLab product docs localization workflow
  • Argo-TOS integration - service for marketing localization workflow
  • Argo-GitLab integration - service handling webhooks
  • Argo GitLab agent - service for preprocessing GitLab product docs markdown files
  • Database & reporting - business analytics and tracking, available within Argo UI

GitLab integration services

Integrations with Language Technology Platforms

AI-powered translation

  • Tech Docs AI-powered translation - Google Cloud Vertex AI with Gemini LLM processing GitLab product documentation, using advanced NLP, chained prompt systems, multiple glossaries and style guide injection, and file transformations and validations
  • Emerging AI tools - standalone AI projects in Claude in the early stages of prototype and trial for translation production, to be transformed into services orchestrated by Argo

Content management system integrations

Supporting tools and services

Central content repository

All localization workflows converge through GitLab as the single source of truth for content and translations. This GitLab-centric approach ensures:

  • Version control for all source and translated content
  • Unified delivery mechanism through Translation MRs
  • Consistent quality gates and approval workflows
  • Integrated CI/CD pipelines for content publishing