Skip to main content
Version: 3.x.x

Usage Tolgee MCP Server

Example Workflow​

Once connected, you can interact with Tolgee in natural language. For example:

You: List all keys in the "homepage" namespace that contain "welcome"

AI: Found 3 keys in namespace "homepage":

  • homepage.welcome_title — "Welcome to our platform"
  • homepage.welcome_subtitle — "Get started in minutes"
  • homepage.welcome_cta — "Sign up free"

You: Translate these keys to German and French using machine translation

AI: Started machine translation batch job. Job ID: 42. Checking status... completed. All 3 keys have been translated to German (de) and French (fr).

Best Practices​

  • Start with read-only commands — Use list_keys, get_translations, or list_projects first to verify that your connection works before making changes.
  • Use a Project API key when working on a single project — This avoids having to specify projectId with every request and limits the blast radius of accidental changes.
  • Be careful with destructive operations — Tools like delete_keys and delete_branch are irreversible. AI assistants are instructed to confirm with you before executing them, but always double-check.
  • Keep tokens secure — Store your API keys in environment variables or a secrets manager. Don't commit them to version control. Use the minimum required scopes and rotate tokens regularly.
  • Leverage Big Meta for better MT quality — If you're using machine translation, store key relationships with store_big_meta so Tolgee can use translations of related keys as context.