Manage localization from your AI tools with Lokalise MCP

The Lokalise MCP Server connects MCP-compatible AI tools — like Claude, ChatGPT, Cursor, Dust and Windsurf — to your Lokalise workspace. It lets you manage localization tasks in natural language instead of switching to the Lokalise dashboard or writing API scripts. MCP (Model Context Protocol) is an open standard developed by Anthropic that gives AI assistants a consistent way to interact with external tools and services.

The Lokalise MCP Server works with any tool that supports the Model Context Protocol. This includes Claude, ChatGPT, Cursor, Windsurf, Claude Desktop, Dust and Claude Code. As MCP adoption grows, any new MCP-compatible client will work with Lokalise automatically.

Most teams complete the setup in under 5 minutes. The MCP Server is cloud-hosted, so there are no packages to install. You add the server URL to your tool's MCP configuration, authenticate with OAuth or an API key, and start using it immediately.

Yes. The Lokalise MCP Server supports OAuth authentication and follows your existing Lokalise role-based access controls (RBAC). Every action the AI assistant performs respects the same permissions your team members have in Lokalise. No separate credentials are needed.

The Project Management toolkit (lokalise_pm) is designed for localization managers. It covers project creation, contributor management, task tracking, glossary management, and file uploads. The Software Development toolkit (lokalise_sd) is designed for developers. It focuses on adding and updating keys, attaching screenshots, checking translation coverage, and downloading files. You can install one or both.

Yes. You can add both toolkits to your MCP configuration. Your AI assistant will choose the right one based on the task you describe.

No. You describe what you need — "create a task for untranslated French keys" — and the AI assistant handles the Lokalise actions for you.