> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tolk.every-day.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Glossary

> Fixed translations of terms, and terms that are never translated.

In the glossary of a project you record how fixed terms are translated in each language, or that they stay untranslated, such as a brand name. AI translation follows the glossary and the editor highlights the terms in the source text.

## Add a term

Click **Add term** (*Term toevoegen*) and fill in:

| Field | Explanation |
| - | - |
| **Term** | The term in the source language |
| **Do not translate** (*Niet vertalen*) | The term stays the same in every language, for example a brand name |
| **Translations** (*Vertalingen*) | The fixed translation per language |
| **Case-sensitive** (*Hoofdlettergevoelig*) | Tolk only recognizes the term with exactly these capitals |
| **Note** (*Notitie*) | Explanation for translators, for example: *niet ‘mand’* |

Deleting a term removes it from the glossary. Existing translations in the editor stay.

## Import from CSV or Excel

With **Import CSV** (*CSV importeren*) you add or update terms from a single CSV or XLSX file for all languages of the project. The first row is a header row, with a column per language:

| Column | Content | Recognized by the header |
| - | - | - |
| The term | The term in the source language (required) | `term`, or the source language (code or name, for example `nl` or `Nederlands`) |
| A target language | The fixed translation in that language | the language code (`de`, `de-DE`, `de_DE`) or the name (`German`, `Duits`, `Deutsch`) |
| Note | Explanation for translators | `note` or `notitie` |
| Do not translate | Yes or no | `do_not_translate` or `niet vertalen` |
| Case-sensitive | Yes or no | `case_sensitive` or `hoofdlettergevoelig` |

Capitals, spaces and hyphens in the headers do not matter. For yes and no, `ja`, `yes`, `true` and `1` work (and their opposites).

```csv glossary.csv theme={null}
nl,de,fr,note,do_not_translate,case_sensitive
Every Day,,,Merknaam,ja,ja
winkelmand,Warenkorb,panier,niet ‘mand’,nee,nee
```

<Tip>
  **Download example file** (*Voorbeeldbestand downloaden*) gives you a CSV with the right header row for this project: the source language, the target languages and the three fixed columns.
</Tip>

### Check the columns

After you choose a file, Tolk shows per column what it recognized: **Term**, **Note**, **Do not translate**, **Case-sensitive**, **Language: …** (*Taal: …*) or **Ignore** (*Negeren*). If something is wrong, change it; the counts update.

* Choose exactly one column for the term.
* Each language and each field can have only one column; with duplicate columns, the first one wins.
* A column with a language that is not in the project is ignored and reported.

Before anything is saved, Tolk shows how many terms are new, how many are updated and which rows are skipped, with the reason.

## How Tolk uses the glossary

* **Glossary table:** with many languages, the term and the actions stay in place while you scroll horizontally. A term with **Do not translate** has a lock icon and appears unchanged in every language column.
* **Editor:** a term in the source text is subtly highlighted, with the fixed translation as a tooltip.
* **AI translation:** the terms that occur in the texts to translate are sent to the AI, including the do-not-translate list. In the **Translate with AI** (*Vertalen met AI*) dialog you see how many terms are included and you can turn the glossary off.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.