@translated/lara-mcp

What it does

This server exposes Lara Translate's professional translation API through the Model Context Protocol. Core functionality includes language detection, text translation with support for context and domain-specific glossaries, and translation memory management. Unlike general-purpose LLMs, Lara uses Translation Language Models (T-LMs) trained on billions of professionally translated segments, delivering translations that preserve industry terminology, cultural nuance, and domain conventions. The translate tool supports multiple output styles (faithful, fluid, or creative), context-aware rendering, and integration with glossaries and translation memories. Additional tools cover language listing, memory lifecycle operations (create, update, delete, import from TMX), and glossary management with CSV import/export.

Who it's for

Teams building multilingual applications or content systems that require professional-grade translation quality beyond what general LLMs provide. Localization engineers managing translation memories and glossaries for large projects. Product teams translating user-facing content while maintaining consistent terminology across deployments. Developers integrating translation capabilities into Claude or other AI workflows via MCP.

Common use cases

• Translate content with domain-specific terminology preserved via shared glossaries
• Automatically detect source language before proceeding with translation
• Build and maintain translation memories (TMX format) for consistent terminology across projects
• Import glossaries and translation memories from existing localization tools (CSV, TMX formats)
• Offer Claude-powered translation within existing applications via MCP integration

Setup pitfalls

• OAuth browser authentication required on first use—the client automatically opens a browser for login. Plan for interactive setup if deploying in headless environments.
• Access Key alternative authentication requires credentials from developers.laratranslate.com. Self-hosted setups need LARA_ACCESS_KEY_ID and LARA_ACCESS_KEY_SECRET environment variables.
• Filesystem read/write access used for TMX and CSV import/export—ensure proper permissions if containerizing or sandboxing.