Skip to content
LokAI Docs

Introduction

What is LokAI?

Section titled “What is LokAI?”

LokAI is an AI-powered localization management platform for software teams. It manages translation keys, glossaries, style guides, and tenant customizations, with LLM-assisted translation workflows at its core.

Unlike traditional localization tools where humans do the translation work and machines assist, LokAI inverts the model: LLMs handle the translation volume while your team focuses on quality, context, and brand voice.

Who is it for?

Section titled “Who is it for?”

| Persona | How LokAI helps | | ------------------------- | -------------------------------------------------------------------------- | | Developers | Add keys via CLI or API, get instant AI translations, integrate with CI/CD | | Product Owners | Enforce brand voice with glossaries and style guides, track coverage | | Localization Managers | Review only what needs attention — AI handles the rest | | Reviewers | Rich context on every translation, feedback loop that improves the AI | | Key Maintainers | Single view of all keys you own, batched context requests |

Key concepts

Section titled “Key concepts”
  • Workspace — top-level container for a company or team
  • Project — a product being localized, with main, target, and secondary languages
  • Translation Key — a translatable string with context, flags, and per-language translations
  • Glossary — workspace-level terminology (preferred terms, forbidden terms, regional overrides)
  • Style Guide — Markdown document instructing the LLM on tone, formality, and conventions
  • Tenant — a B2B customer with their own terminology customizations, resolved by the LLM at translation time
  • Access Grant — a role assignment scoped to the workspace or a project, optionally narrowed by language

How it works

Section titled “How it works”
  1. Push keys — developers push translation keys via the CLI or API, with source text and context
  2. AI translates — the LLM translates into all target languages, applying the glossary and style guide
  3. Review — confidence scores highlight which translations need human review
  4. Approve & export — approved translations are pulled back into the codebase via CLI or CI/CD

Next steps

Section titled “Next steps”