Tolgee AI Translator is designed around one idea: machine translation is far more accurate when the model understands context. Tolgee's SDKs and integrations automatically extract context from your app, so the AI sees surrounding keys, screenshots, and the UI situation around each string instead of guessing. You can further provide a project description, language-specific notes, and key-level descriptions to guide tone of voice and domain terminology.
Beyond out-of-the-box quality, the feature is highly controllable. Prompt customization lets you edit Tolgee's default prompt fragments and write your own using variables that expose key data, screenshots, and context. An AI Playground lets you test prompts against your real translation data, compare different LLM models, and save the best configuration as default. You can also connect your own LLM providers such as OpenAI, Anthropic, or Google Gemini — including self-hosted models — with load balancing to avoid rate limits. A newer MCP server connects Tolgee to AI coding assistants like Claude Code, Cursor, and Codex so translations can be generated and updated directly from the editor.
Tolgee itself is open source and can be self-hosted, with a cloud offering on top. Pricing uses unlimited seats and is based on hosted words and MT credits: a free tier covers small projects (30,000 words, 3 seats), while paid cloud plans — Translate, Optimise, and Manage — scale by word volume and advanced features, with Enterprise and on-premises options for large organizations. Tolgee holds ISO 27001 and GDPR compliance positions.