--- title: "Create a glossary to improve terminology accuracy" description: "Define misheard-to-correct term pairs, import terms in bulk from a CSV, and apply the glossary when you create a transcription" canonical_url: "https://docs.ardge.com/en/knowledge/scribestt/glossary" --- # Create a glossary to improve terminology accuracy > Define misheard-to-correct term pairs, import terms in bulk from a CSV, and apply the glossary when you create a transcription ## Overview Company names, product names, and technical terms are where speech recognition goes wrong most often, and the same word can come out spelled differently in different passages. Defining term mappings in a glossary reduces those errors and keeps a term consistent across the whole transcript. Each glossary entry pairs a **Misheard term** with a **Correct term**: the misheard term is the wrong form recognition tends to produce, and the correct term is what should appear in the transcript. This article covers creating a glossary, importing terms in bulk from a CSV, and applying it when you create a transcription. ## What you'll learn - Create a glossary and define misheard-to-correct term pairs - Import terms in bulk from a CSV file, appending to or overwriting what's there - Apply a glossary when you create a transcription **Estimated time: about 10 minutes** ## Applies to - ARX OS 1.0 and later - ScribeSTT 1.3.0 and later ## Before you begin - You have a list of terms to correct, each with the form that tends to be misrecognized and the correct form ## Step 1: Create the glossary 1. In the **Workspace** group in the left navigation, go to the **Glossaries** page 2. Click **New** at the top right to open the **New glossary** dialog 3. Fill in the fields as follows
Field Value Rules and notes
Glossary name A name that identifies this glossary Required
Glossary language The language this glossary applies to Optional. Chinese ( zh ), English ( en ), Japanese ( ja ), or Korean ( ko )
Terms The misheard-to-correct mappings Required, at least one. Neither column can be left empty
1. Fill in the first entry in the **Terms** table: **Misheard term** on the left, **Correct term** on the right 2. Click the add icon at the right of the header to add a row, or the remove icon on a row to delete it 3. Click **Create** **Both columns are required:** If any row has only one column filled in, submitting shows "Field can't be empty"; with no terms at all, it shows "Add at least one term". The page returns to **Glossaries**, where the new glossary appears as a card in the list. ![The empty Glossary Library](https://docs.ardge.com/screenshots/knowledge/en/scribestt/glossary/step-1-1.webp) ![Add Glossary and Import .csv](https://docs.ardge.com/screenshots/knowledge/en/scribestt/glossary/step-1-2.webp) ![Entering fuzzy and correct words](https://docs.ardge.com/screenshots/knowledge/en/scribestt/glossary/step-1-3.webp) ![The new glossary](https://docs.ardge.com/screenshots/knowledge/en/scribestt/glossary/step-1-4.webp) ## Step 2: Import terms in bulk from a CSV With a lot of terms, import a CSV instead of typing row by row. 1. Prepare a CSV file, one entry per line, formatted as "misheard term, correct term":```csv misheard term 1, correct term 1 misheard term 2, correct term 2 misheard term 3, correct term 3 ``` 2. On the **Glossaries** page, click the glossary's card to open its detail page 3. Click **Import CSV** at the right of **Term list**, and choose one of the import modes:
Import mode Result
Append Keeps existing terms and merges the CSV in. Where a misheard term appears in both, the CSV wins
Overwrite Clears existing terms and replaces them with the CSV
1. Select the CSV file 2. Click **Save** at the top right of the page **Importing doesn't save:** The import only updates the **Term list** on screen. Leave the page without clicking **Save** and the imported content is lost. **Term list** shows how many terms are there after the import, with the total below the table. **You can import at creation time too:** The **Terms** section of the **New glossary** dialog also offers **Import CSV**. An import there replaces the dialog's current table outright — there's no append or overwrite choice. ## Step 3: Apply the glossary when creating a transcription 1. Go to the **Transcriptions** page and click **New** 2. In the **Create new transcription** dialog, select your files and the **Output formats** 3. Pick the glossary to apply in the **Glossary** dropdown 4. Click **Create** After submitting, that record's **Glossary** column in the **Transcriptions** list shows the glossary's name; click it to open the glossary. ![Applying a glossary when creating a transcript](https://docs.ardge.com/screenshots/knowledge/en/scribestt/glossary/step-3.webp) **One glossary per transcription:** To cover several sets of terms at once, merge them into a single glossary. ## Editing and deleting a glossary On the glossary detail page you can edit **Glossary name**, **Glossary language**, and the **Term list** directly, then click **Save**. Click **Delete** and confirm in the "Confirm delete" dialog to remove the glossary. The page shows "Glossary " " deleted successfully". Deleting a glossary doesn't delete the transcriptions and tasks that used it — those records stay, they just no longer carry a glossary association. ## Next steps With the glossary created, apply it every time you create a transcription. For the full upload and review flow, see [Upload an audio or video file and get a transcript](https://docs.ardge.com/en/knowledge/scribestt/upload-and-transcribe). ## Related articles - [Upload an audio or video file and get a transcript](https://docs.ardge.com/en/knowledge/scribestt/upload-and-transcribe) ## Sitemap See the full [sitemap](https://docs.ardge.com/sitemap.md) for all pages.