The control center
for AI localization

AI-powered localization uses large language models (LLMs) from OpenAI, Anthropic, Google, and others to translate software strings with contextual awareness. Unlike traditional machine translation (MT) that operates on isolated sentences, AI models can process metadata — project descriptions, key-level context, tone instructions, and character constraints — to produce translations that fit your product naturally.

SimpleLocalize bridges the gap between raw AI capability and production-ready localization by acting as a context layer and caching system between your codebase and the AI model. You define the context once, and every translation request inherits it automatically. Combined with smart caching that only re-translates changed keys, AI localization becomes both cost-effective and consistent.

The Context Engine sends structured metadata alongside every translation request. This includes the project description (e.g., “a fintech app for Gen Z”), key-level descriptions and screenshots, character limits, placeholder rules, and tone instructions (formal vs. informal).

This context prevents common AI translation errors — like translating “Home” as a building instead of a navigation button — without requiring you to write custom prompts for every string. Read more about how context improves AI translations.

OpenRouter is a unified API gateway that gives you access to dozens of AI models — Gemini, Claude, OpenAI, Llama, Mistral, and more — through a single endpoint. SimpleLocalize integrates with OpenRouter so you can switch between models without changing your workflow.

This means you can use fast, cheap models from Google for bulk translation of UI strings, and switch to Claude for nuanced marketing copy — all within the same project. You can also connect local models via Ollama for air-gapped, privacy-first translation workflows.

Machine translation (MT) services like DeepL and Google Translate use specialized neural networks trained specifically for translation. They are fast, predictable, and cost-effective for straightforward content. AI translations using LLMs offer deeper contextual understanding — they can follow tone instructions, respect character limits, and adapt to project-specific terminology.

In practice, the best approach is often hybrid: use MT as a fast baseline, then apply AI where context matters. SimpleLocalize supports both workflows — compare DeepL, Google Translate, and OpenAI to decide what fits your project.

MCP (Model Context Protocol) is an open standard that allows AI agents to interact with external tools and data sources. SimpleLocalize provides an official MCP server that AI coding assistants — in Cursor, Windsurf, Claude Desktop, or any MCP-compatible client — can use to query, update, and translate localization data directly from the IDE.

This enables workflows where an AI agent can find missing translations, add new keys, and trigger context-aware AI translation without the developer switching tools. It's the bridge between AI-driven development and production-ready localization.

No. AI translations are available to all users regardless of plan. You can provide your own OpenAI or OpenRouter API key and pay only for the tokens you consume. SimpleLocalize does not charge additional fees on top of the provider's pricing. For DeepL and Google Translate, a monthly character allowance is included in paid plans with unused characters rolling over.

SimpleLocalize supports OpenAI models, and any model available through OpenRouter — including Gemini, Claude, Llama, Mistral, and dozens more. You can also connect local models via Ollama or any OpenAI-compatible endpoint. Additionally, DeepL and Google Translate are available as machine translation providers.

SimpleLocalize tracks which keys have been translated and which have changed since the last run. When you trigger AI translation, only new or modified keys are sent to the model — the rest are served from the translation memory cache. For a project with 2,000 keys where only 10 changed, you pay for 10 translations instead of 2,000. Use our auto-translation cost calculator to estimate savings.

Yes. SimpleLocalize supports any OpenAI-compatible API endpoint. You can run models locally using Ollama (e.g., Llama, Mistral, Gemma) and point SimpleLocalize to your local endpoint. Your translation data never leaves your infrastructure — ideal for teams with strict data-privacy or regulatory requirements.

The SimpleLocalize MCP server allows AI coding assistants (Cursor, Windsurf, Claude Desktop) to interact with your localization data directly from the IDE. Install the MCP extension, connect it to your SimpleLocalize project, and your AI agent can query missing translations, push new keys, and trigger auto-translation without switching tools.