Skip to content

Glossary Management

Last updated: Aug 13, 2025

OmniTranslate's glossary, or "Context," is a powerful tool that gives you fine-grained control over your translations, ensuring consistency and accuracy across all your projects. This guide covers the advanced features of the glossary system.

Glossary Section

Anatomy of a Term

Anatomy of a Term

Each term in your glossary has several components you can configure:

  • Raw: The original, untranslated word or phrase from the source text (e.g., 叶安平).

  • Translation: The English translation you want the AI to use (e.g., Ye Anbing). This demonstrates a custom translation, as the direct pinyin for 叶安平 is "Ye Anping". If you leave this field blank, the AI will be instructed to ignore this term and not translate it.

  • Description: An optional field where you can provide specific instructions to the AI about how to handle the term. This is a powerful feature for guiding the translation. For example:

    • For a term with multiple possible translations: "Use 'Your Majesty' when addressing the emperor, and 'His Majesty' otherwise."
    • To preserve a specific style: "Keep this term in all lowercase."
  • Gender: You can set a gender (Masculine, Feminine, or Neuter) for a term. This helps the AI use the correct pronouns (he/she/it) when referring to the term in the translation, which is especially useful for character names.

  • Important: Marking a term as "Important" gives it a higher priority in the translation process. This makes it more likely that the AI will use your specified translation, even in complex sentences.

System vs. User Terms

System vs User Terms

  • System Terms: These are terms that OmniTranslate's AI automatically extracts from the text while translating. They are provided as suggestions and may not always be perfect.
  • User Terms: These are terms that you have manually added, edited, or confirmed. User terms always take precedence over system terms.

By managing your glossary effectively, you can significantly improve the quality and consistency of your translations, creating a more enjoyable reading experience.

WARNING

Currently, OmniTranslate treats glossary terms as strong suggestions rather than strict rules. The model may still choose to use other variations of a term based on context, especially if the provided translation significantly differs from the original.

Bulk Management with CSV

CSV Format

For users with large glossaries, OmniTranslate supports adding and updating terms in bulk using a CSV file. For a step-by-step tutorial, see our How to Bulk Import and Export Terms guide.

CSV Format

Your CSV file should have three columns: raw, translation, and description.

csv
raw,translation,description
白小纯,Bai Xiaochun,Main character #masculine #important
周紫陌,Zhou Zimo,Love interest #feminine
灵溪宗,Spirit River Sect,The main character's sect

Column Rules:

  • raw → The original term in the source language.
  • translation → The translated term. Leave this empty to disable the term.
  • description → A short description of the term. You can also include special tags here.

Special Tags

Tags are written inside the description column and control how the term is treated:

  • #masculine or #male: Sets the gender to Masculine.
  • #feminine or #female: Sets the gender to Feminine.
  • #important: Marks the term as Important.

If no gender tag is provided, the term will default to Neuter.

Description vs Tags

Description vs Tags

  • Any text in the description column without a # is treated as plain description.
  • Words or phrases starting with # are recognized as tags (properties).
  • You can include both a description and tags in the same cell.

Example:

Main character #masculine #important

Here:

  • Description: Main character
  • Tags: #masculine, #important

Backwards Compatibility

The bulk import feature is also compatible with the older, three-column format: raw,translation,gender. If you upload a file in this format, the system will automatically detect it and process it correctly.

NOTE

This is because the previous version of the import feature used a gender column instead of description.*