Improving translations with the glossary

Emfas automatically translates your product content using AI, delivering high-quality translations that work well for most use cases. However, when you have specific terminology - like technical terms, brand names, or preferred phrasings - you can use the translation glossary to ensure these terms are always translated exactly the way you want.

When to use the translation glossary

The translation glossary is perfect for:

• Technical terminology: Industry-specific terms that need precise translation (e.g., "organic cotton" → specific textile industry term)
• Product-specific language: Terms unique to your products or brand that should be translated consistently
• Preferred translations: When you have a specific preference for how something should be translated, even if multiple options are valid
• Fine-tuning edge cases: Occasional terms where the AI makes an unconventional choice that you want to correct

Accessing the glossary

To access your translation glossary:

1. Click Translations in the left sidebar
2. Click Glossary at the top of the page

This opens the glossary table where you can view and manage all your translation terms.

Understanding the glossary table

The glossary is organized as a table with your base language terms in the first column, and each translatable language in subsequent columns:

• First column: Your base language term (the term as it appears in your original content)
• Language columns: Translation for each language you've configured in Emfas
• Each row represents one term and all its translations

Adding glossary terms

To add a new term to your glossary:

1. Click the Add term button
2. Enter the term in your base language
3. For each language column, enter the corresponding translation
4. Press Enter or click away to save

Editing terms

To edit an existing term or translation:

1. Click directly on the text you want to edit
2. Make your changes
3. Press Enter or click away to save

Removing a translation

To remove a translation for a specific language while keeping the term:

1. Click on the translation you want to remove
2. Delete all the text
3. Press Enter to clear the translation

The base language term and other translations remain intact.

Deleting terms

To completely remove a term and all its translations:

1. Hover over the term row
2. Click the X button that appears
3. The term and all its translations are deleted

How glossary terms are applied

Once you've added terms to your glossary, they're automatically used when generating translations:

• When you generate AI translations, Emfas checks your glossary for the target language
• Any terms found in your glossary are provided to the AI to ensure consistent translation
• The AI incorporates these terms naturally into the translated content

You don't need to do anything special - glossary terms are automatically applied whenever you generate translations for languages where you've defined terms.

Best practices

Start with problematic terms

Don't try to build a comprehensive glossary upfront. Instead, add terms as you notice the AI making translations you want to adjust. This keeps your glossary focused and manageable.

Keep terms specific

Add specific terms or phrases rather than generic words. For example:

• ✅ Good: "organic cotton jersey" → specific textile term
• ❌ Less useful: "cotton" → common word that translates fine

Use the glossary for consistency

If you notice the AI translating the same term differently across products, add it to your glossary to ensure consistency across your entire catalog.

Review generated translations

Even with a glossary, it's good practice to review your first few translations to ensure the terms are being applied correctly and producing natural-sounding results.

After adding glossary terms

Once you've set up your glossary:

1. Generate new translations - The glossary terms will be automatically used
2. Review the results to confirm the terms are being applied as expected
3. Add more terms to your glossary as needed when you encounter new edge cases