Skip to documentation
DocumentationThe workspace

AI translation and dictionaries

Use project context, shared terminology, and connected translation providers to prepare drafts for review.

On this page

Connect a translation provider

ReRune integrates with DeepL and OpenRouter. Configure the integration with the provider credentials required by your account. Where project AI settings are available, choose the provider and model you want the project to use.

The integration uses the connected provider account. Provider availability, supported languages, model behavior, and provider charges depend on that service and configuration. Check the provider response if a translation request fails.

Maintain shared terminology

Use dictionaries for terms that need consistent treatment across messages. Record the product term and its intended translation. Include brand names and domain-specific words that a general-purpose model might interpret incorrectly.

Pair dictionary entries with project description and translation guidelines. A glossary explains terminology; project context explains the audience, the product, and the desired voice. Review both when changing your product language.

Review generated drafts

  1. Start with accurate main-language text and enough context to resolve ambiguity.
  2. Select the messages and target languages and run the available translation action.
  3. Check the result for meaning, terminology, placeholders, plural forms, and space constraints.
  4. Save the reviewed text and publish through your normal delivery workflow.

Treat generated output as a draft. A fluent sentence can still use the wrong domain term or change the meaning of a short action label.

AI translation, context, and draft review

Follow a real OpenRouter translation request, inspect the generated draft, make an editorial correction, and publish it. The example uses synthetic reading-app content.

Watch on YouTube